International Journal of Production Research,Vol. 45, No. 11.docx

International Journal of Production Research,
Vol. 45, No. 11, 1 June 2007, 2567–2594
Collective customer collaboration impacts on supply-chain
performance
GREG ELOFSONy and WILLIAM N. ROBINSON*z
yFisher Graduate School of International Business Monterey
Institute of International Studies,
460 Pierce Street, Monterey, California 93940, USA
zRobinson College of Business, Computer Information Systems
Department, Georgia State
University, Atlanta, GA 30302, USA
(Revision received November 2006)
Facilitated by a Collective Customer Collaboration (C3) system,
the impact of
customer knowledge and demand information on supply-chain
performance is
examined. The basic characteristics and theoretical benefits of
C3 systems are
presented and tested using a multi-agent supply-chain
simulation. The simulation
test results indicate that (1) supply chains employing C3 front-
ends, without
sharing information, show performance improvements over
traditional supply

chains; (2) supply chains employing C3 front-ends, while also
sharing information,
show performance improvements over those that did not share
information; and
(3) centralized supply chains employing C3 systems, while also
sharing informa-
tion, show performance improvements over their decentralized
counter-parts.
Keywords: Collective Customer Collaboration (C3) system;
Supply chain;
Customer knowledge
1. Introduction
The demand for co-designed products, where customers
participate in sharing
knowledge and information to specify products which have
custom characteristics, is
continually increasing (Piller et al. 2005). And, recently, some
have asserted that
having the interaction skills and ability to match custom
configurations to the needs
of customers during a process of co-design is becoming a
primary source of
competitive advantage (Sheth et al. 2000, Seybold et al. 2001).
Further, while B2B
markets have seen various forms of customization become
relatively more
commonplace (Homburg et al. 2000), co-design in most
consumer markets is still
in a nascent stage, and therefore may afford rich research
opportunities for
exploration.
Collective Customer Collaboration (C3) is a recent step in co-

design, and entails a
group of customers taking part in product design as well as
committing to product
purchase (Piller et al. 2005, Ogawa and Piller 2006). Industry
examples of C3
approaches include those undertaken by Threadless.com, Muji,
and Yamaha.
Threadless has a community of customers that submit, inspect,
and approve new
t-shirt graphic designs; Muji uses an online customer
community to help design
*Corresponding author. Email: [email protected]
International Journal of Production Research
ISSN 0020–7543 print/ISSN 1366–588X online � 2007 Taylor
& Francis
http://www.tandf.co.uk/journals
DOI: 10.1080/00207540601020528
consumer products such as bean-bag sofas, bookshelves, and
portable lamps; and
Yamaha used a team of users to test and make recommendations
for musical
instruments, and has teamed with online C3 providers in the
process.
As a variant of mass customization, C3 entails, to some extent,
considerations of
supply-chain redesign (Boynton and Victor 1991, Pine 1993,
Pillar et al. 2004,
Salvador et al. 2002, Westbrook and Williamson 1995). Also,
because C3 promises

particularly high levels of customization, the redesign needed
may be significant.
In the context of Knowledge and Information Technology
Management systems
(KITM) in manufacturing, our interest in this paper is to
examine the basic
characteristics of C3 systems, and also to evaluate the effects of
C3-generated
customer knowledge and demand information on supply-chain
efficiencies over a
variety of designs. In line with testing the predictions made by
Piller (Piller et al. 2005)
and the work done by Aviv (2001), Cheung and Lee (2002),
Kulp (2002) and
Moinzadeh (2002), we are interested in clarifying how
information sharing and
coordination via a C3 system affects production costs. We
believe these to be
interesting issues, insofar as the theories about C3-based
savings have yet to be
broadly tested. And, the feasibility of using C3 in various
supply-chain designs is yet to
be explored – for example, those designs being dependent on
factors such as anti-trust
regulations and competitive security considerations (Lee and
Whang 2000).
To this end, we discuss the basic principles of C3, and use a C3
system exemplar
as cited by Ogawa and Piller (2006) to illustrate its interaction
with three, separate
supply-chain configurations. We use a multi-agent simulation,
benchmarked against
the analytical model derived by Anand and Mendelson (1998),
to evaluate the

following questions: (1) do supply chains employing C3 without
sharing information
show performance improvements over traditional supply chains,
as lead times
diminish? (2) Do supply chains employing C3, while also
sharing its information,
show performance improvements over their non-sharing
counter-parts, as lead times
diminish? (3) Do wholly owned or centralized supply chains
employing C3, while
also sharing its full information, show performance
improvements over decentralized
supply chains, as lead times diminish?
This article is organized as follows: in section 2, we describe
the principles and
theories of C3 systems, and provide the literature and
background for our three
research questions. In section 3, we detail the components,
assumptions, and design
of our simulation model. In section 4, we describe our research
methodology. In
section 5, we provide the simulation results. In section 6, we
discuss the results in
light of the predictions found in the literature and compare them
to our hypotheses.
Finally, section 7 presents our conclusions.
2. C3 system characteristics and assertions
Recently, Piller et al. (2005) and Ogawa and Piller (2006) have
proposed and refined
the idea of Collective Customer Collaboration to characterize
the process of using
customer knowledge in an effort to take the best features of
mass customization and

postponement, and include collaborative customer knowledge to
create the ‘‘most’’
custom products at prices that reflect mass-production
manufacturing. In general,
C3 is an outgrowth of co-design, where co-design is
characterized as a process of
collaborative value creation between multiple actors (Wikström
1996, Ramirez 1999,
2568 G. Elofson and W. N. Robinson
Prahalad and Ramaswamy 2004, Berger et al. 2005). Through
basic co-design,
customers detail their product specifications and engage in
product definition
processes by translating their requirements into the physical
domain (von Hippel
1998, Khalid and Helander 2003). Further, communities for co-
design reflect expert
knowledge of customer groups that interact not only with one
company, but
importantly, also with each other. Communities for co-design
extend beyond
organizational structures, spanning functional boundaries to
create common
knowledge and value (Gibbert et al. 2002). C3 necessarily
entails an interplay
between customer and manufacturer. Proponents of this
approach assert that it
offers access to a rich stream of information that allows the firm
to cut back on fixed
costs that came about due to the necessity of maintaining a high
level of operational
flexibility. This cost-saving potential is founded on the

capability to design and
re-design the routines to utilize co-design information and to
combine this infor-
mation with existing knowledge to improve the efficiency of the
whole value system
(De Meyer et al. 2002, Rowley 2002, Thomke and von Hippel
2002). Further, this
challenge includes utilizing the information from the co-design
process to enhance
the loyalty of customers towards the firm (Pine et al. 1995,
Berger et al. 2005).
Proponents of C3 argue that there are both ‘‘soft’’ and ‘‘hard’’
advantages over
extant customization approaches. These include the following
advantages:
1. Mass customization provides too many choices and excess
variety, while C3
allows one to specify instead of choose.
2. Postponement can result in inventory waste due to product
flops, while C3
begins the full manufacturing cycle only after customers show
their full
commitment to purchase a specific item.
3. C3 approaches promise greater support for creative activities
(von Hippel and
Tyre 1995, Sawhney and Prandelli 2000, Nemiro 2001, Gascó-
Hernández and
Torres-Coronas 2004) and increases customer integration into
company
activities (Schubert 2000).
4. C3 enables manufacturers to avoid having to make risky

decisions about what
components to prefabricate or about the optimal timing of
postponement.
5. C3 does not require interactions with individual customers
nor does it require
running manufacturing lots of size one.
Overall, C3 holds three major advantages in the course of
customer co-design: (1) the
generation of customer knowledge to provide a better starting
(pre-) configuration;
(2) the support of collaborative co-design fostering joint
creativity and problem
solving; and (3) the building of trust and the reduction of the
perception of risk
(Berger et al. 2005).
2.1 Exemplar C3 system
Ogawa and Piller (2006) cite the CMP broker (Elofson and
Robinson 1998) as an
exemplar C3 system, and we use this system to generate our
simulation assumptions.
The CMP broker uses the World Wide Web as a channel through
which knowledge
of individual buyers is represented in a multi-stage bargaining
action. The bargaining
action takes place in three stages: (1) knowledge of like-minded
buyers is used to
group them through a collaborative filtering algorithm; (2)
product requirements
from the like-minded buyers are clarified and agreed upon
through an automated
Supply-chain performance 2569

negotiation session; and (3) the product specification and the
quantity of orders is
put out to suppliers for bid.
An example of how the CMP broker is used is shown in figure
1. In the figure,
customers A and B are members of some n number of like-
minded buyers who are
seeking to buy a personal computer that primarily supports
surfing on the Web.
Customer B wants lots of CPU power, a large flat-screen
monitor, and substantial
storage capacity. Customer A wants a large flat-screen monitor,
modest CPU power
to save money, and modest storage capacity. Together,
customers A and B – along
with the other like-minded buyers – engage in a negotiation
over the sought-after
PC’s product specification until it is agreed that they will try to
buy a PC with a large
flat-screen monitor, modest CPU power, and large storage
capacity. Once this
agreement has taken place, the PC specification, along with the
number of orders, is
submitted to one or more worldwide suppliers for pricing and a
deal is made.
The goal of a CMP broker, therefore, is to facilitate the process
of providing
buyers with products that are better suited to their needs, but
without the penalty of
higher prices – that is, to consolidate buyer-purchasing
decisions and achieve better

results through lower search costs and higher bargaining power.
2.2 C3 system architecture
Figure 2 illustrates the logical component architecture of the
CMP broker. Each of
the rectangles represents a computerized component within the
UML interaction
diagram. Customer, supplier, and broker are the three main
components; the broker
consists of three subcomponents. The shadowed rectangles of
customer and supplier
indicate that multiple components (e.g. multiple customer
systems) can interact with
the broker.
A variety of physical architectures can realize the logical
architecture. Typically,
the three main components each reside on separate computers.
However, each
Bob’s spec:
inexpensive,
large flat-screen monitor,
modest CPU,
modest storage capacity
CMP Broker
Assisted collective
contract negotiation
Assisted collective
product negotiation
Final group contract:

modest CPU,
large storage capacity
Negotiated group spec:
modest CPU,
Carol’s spec:
powerful CPU,
Figure 1. An illustration of an CMP broker.
broker subcomponent can also reside on separate computers.
Moreover, the broker
can be partitioned into the common layers of presentation,
application, and
database, each of which can be placed on separate computers,
thereby defining a
four-tier architecture (in which the customer or supplier
occupies the fourth client-
presentation tier).
There are a number of implementations for each of the
components – perhaps, of
the most interest are the collaborative filtering systems
(Resnick and Varian 1997,
Herlocker et al. 2004) and the negotiations systems used to
generate product

specifications (Elofson and Robinson 1998, Robinson and
Volkov 1998, Su et al.
2001, Jin Baek and Segev 2003, Dang and Huhns 2005, Hyung
Rim et al. 2005).
A sequence of the steps that a CMP broker must take to create a
custom order
is presented next. The following numbered steps refer to the
sequence numbers
in figure 2.
1. Start: a buyer states a request for a PC.
(1.1) Preference acquisition: buyer specifies PC preferences,
usage informa-
tion, and price preference. (Many other buyers contribute their
buying
preferences.)
(1.2) Collaborative filtering module uses knowledge of
preferences to find
group(s) of like-minded buyers.
2. Based on buyer group characteristics, the negotiation module
suggests
existing product(s) as a straw-man from which to begin
customization (e.g.
Pentium 3, 1 Gigabyte Hard drive, 17" Flat screen, DVD).
3. Buyer feedback: buyer requests deal improvement mediated
by the
negotiation module.
Customer
Broker

Supplier
Request
processor
Collaborative
filter
Specification
generator
collaborative preferences
preferences
1 request product
2 collaborative spec
3 update preferences
4 updated spec
5 accept spec
6 accept bid
5.1 product spec
5.2 bid
6.1 accept bid
Figure 2. Component architecture of the CMP broker (UML
2.0).

4. Contract restructuring: negotiation module restructures the
deal to create new
contracts for buyers to examine.
. Iterative restructuring: repeat steps 3–4.
5. Proposal submission: buyer requests PC deal.
(5.1) Auction module offers custom product to suppliers for bid.
(5.2) Basic matching: suppliers provide bids, which are matched
with
preferences.
6. Deal creation: buyers and suppliers agree to deal: end.
Thus, the main activities that are mediated by the CMP broker
include: (1) a buyer
considers a product purchase; (2) the CMP broker finds other
people interested in
making a similar purchase; (3) a negotiation session among
buyers to agree on
a specific set of product characteristics takes place; and (4) the
custom product
specification goes out for bid. The result of such a session is a
completed
specification of n number of PC orders to be filled by a given
supplier.
The product configuration specification is the planning input for
the customized

manufacturing, assembly, and delivery steps. While these
processes are cost drivers
Piller et al. (2004) assert that if manufacturing and assembly are
performed
on-demand instead of on-stock, savings may nevertheless be
possible, including
the following:
1. Inventory reduction or elimination of inventory in
distribution chain;
reduction in safety stock.
2. Planning complexity reduction; reduction of adaptation costs
(of planning
decisions), fashion risk, and development costs (production
flops).
3. Capacity utilization and stability: no bull-whip effect via
stable processes and
the reduction of the over-capacity required in make-to-stock
systems to adopt
to short-term changes in trends.
Of particular interest in our study is the impact of the
configuration information
process, which can reduce supply-chain inventory (assertion 1
above). However, we
assume that distinguishing information transfer as simply
sharing versus not sharing
information is insufficiently nuanced for evaluating supply-
chain designs. For
example Lee and Whang (2000) point out that supply-chain
partners must take care
when sharing sensitive cost data, such as production yields or
parts pricing – sharing
full information can lead to less bargaining power and falling

profits. Further,
confidentiality must be a top priority under some competitive
circumstances, and, to
further complicate issues, information sharing can lead to a
violation of antitrust
regulations. Therefore, we wish to investigate information
transfer under specified
circumstances, to reflect the dilemmas that face supply-chain
designers.
3. C3 interaction with supply chain
For our evaluation of the interaction of the C3 system with a
supply chain,
we assume that the supply chain under consideration has
exclusive use of the
CMP broker, and thus has exclusive C3 production specification
information.
Consistent with current C3 systems (Piller et al. 2005, Ogawa
and Piller 2006), we
assume that the supply chain manufactures the products, rather
than sending the
specification to auction (steps 5.1 and 5.2 of section 2.2).
Therefore, the supply-chain
configurations evaluate using a make-to-order production policy
of C3 products.
However, the information kind and transfer varies: demand-
forecasting methods
vary (traditional vs. CMP based) and information transfer varies
(sharing and not
sharing among nodes).

Make-to-order strategies are typically used when there are long
lead times that
allow orders to begin upstream in the supply chain, and
customers’ needs are not
excessively time sensitive. For example Lee (1996) cites a disc-
drive manufacturer
that receives different orders from OEMs. They assume a
supplier of raw materials,
a manufacturer that supplies parts, and a manufacturer that
produces units in two
stages within time T. The first stage of production generates the
generic unit in time t,
and the generic stock units are placed in inventory. To
differentiate the generic
products and personalize them, they are put through a second
process in a time T � t.
After time T, the products may be shipped to either a
distribution centre or buyer.
A C3’s CMP broker improves on this strategy by enabling an
increase in order size as
well as improved information precision in demand forecasting
(Robinson and
Elofson 2001).
Given that our C3 system increases order size and information
precision, we are
interested in two supply-chain variables: (1) information
sharing; and (2)
centralization, in the context of centralized and decentralized
control.
3.1 Information sharing
Information sharing has been treated broadly in the literature,
and has included

online supplier information with buyer inventory levels
(Moinzadeh 2002), vendor
managed inventory (VMI) information (Cheung and Lee 2002,
Kulp 2002), non-
specific advance demand information (Thonemann 2002),
continuity of flow
(Heikkila 2002), warehouse inventory control effects (Qi-Ming
et al. 2002) and
shared forecasts (Aviv 2001). Notwithstanding the many
treatments of information
sharing in the literature, some generalizations may be made
from the results.
First, in keeping with Milgrom and Roberts’ (1990) assertion
that information
may substitute for inventory in economic terms, the literature
on information
sharing shows positive, albeit varying, levels of savings (Chen
1998, Cachon and
Fisher 2000, Lee et al. 2000, Moinzadeh 2002).
Second, information sharing has positive savings for a variety
of conditions: these
include VMI programs (Kulp 2002), collaborative forecasting
(Aviv 2001), and
online order information (Moinzadeh 2002). However, reports
of when information
sharing is most effective vary. Lee et al. (2000) showed that
benefits were greatest
when demand was significantly correlated over time, while
Moinzadeh (2002)
reported that savings were greatest under extraordinary
circumstance such as very
large orders or very low inventory.
Additionally, concerning the information which would be

shared, information
precision has typically shown positive results for inventory
savings Anand and
Mendelson (1998) and Kulp (2002). Not surprisingly, these
studies showed that
precision in forecasts and orders resulted in inventory savings,
especially under
conditions of delayed differentiation.
Overall, information sharing has positive results on inventory
savings. Therefore,
we would expect that technologies that facilitate the sharing of
information, together
with increasing the precision of information, would also help to
increase inventory
savings. Thus, our first assertion is as follows:
Assertion 1: as information sharing increases across a supply
chain, inventory
savings increase.
3.2 Centralized control
Centralized control exists when decisions on how much, and
when, to produce goods
are made centrally, based on material and demand status of the
entire system.
Decentralized control indicates situations where each individual
unit in the supply
chain makes these decisions based on local information, and has
the decision
authority to do so (Lee and Billington 1993). Even though we

might associate
centralized control with a wholly owned supply chain, and
decentralized control with
a multiply-owned supply chain, this is not always the case. Lee
and Billington (1993)
point out that ‘‘. . . many companies have intentionally
decentralised control of their
business units or functions’’ (p. 835).
These control considerations have received some significant
attention in the
literature. Anand and Mendelson (1997), early on, studied
various levels of control
and concluded that the issue was highly nuanced and contingent
on a multiplicity of
factors (decision rights, information location, supply-chain
design) deserving greater
study. Chen (1999) found that decentralized supply chains have
optimal replenish-
ment strategies when the upstream suppliers are given customer
demand information
beyond downstream orders. Kouvelis and Gutierrez (1997) and
Lee and Whang
(1999) found that centralized control showed greater
profitability, especially when
there is significant fluctuation in demand.
The overall thrust of the literature suggests that centralized
control, with its full
disclosure and high coordination, is generally advantageous for
overall supply-chain
performance. Also, we have seen no evidence in the literature
that indicates that
greater information sharing confounds the other characteristics
of centralized
control (that is, having more precise information does not

interfere with
coordination or material flows). Consequently, we suggest the
following:
Assertion 2: centralized control shows greater overall supply-
chain performance
than decentralized control, and the performance differences will
increase as
delivery windows decrease.
3.3 Information precision and batch size
Within the bounds of supply-chain management, a C3 system
changes the
characteristics of information precision. We assume that,
because a C3 system
helps define product configurations that are agreed upon by a
group of individuals,
individual make-to-order requests may instead become batched,
make-to-order
requests. We assume that information precision in demand
forecasting is increased
insofar as large-scale orders are made through buyer
consolidation and early
purchase commitment, and so forecasting for non-stock items is
obviated with the
presence of a C3 system.
Cachon (1999) showed that retail orders increasing in batch size
resulted in lower
inventory costs for suppliers. Therefore, because C3 aggregates

orders we expect that
the opportunity for lower inventory costs exist. Further Kulp
(2002) and Anand and
Mendelson (1998) showed increased savings with information
precision in demand
forecasts. Cvsa and Gilbert (2002) illustrated that with
sufficient demand
information certainty, advanced purchasing further increases the
possibility for
strategic market leadership. Consequently, we make the
following assertions
regarding the C3 system used in our simulations:
Assertion 3: a C3 system’s batching capability will likely lower
inventory-holding
costs.
Assertion 4: a C3 system’s increased information precision will
likely lower
inventory-holding costs.
These four assertions lead us to the following hypotheses:
Hypothesis 1: a supplier using C3 systems will have decreased
inventory-holding
costs compared to a supplier with no C3 system.
We expect this because the supplier with a C3 system will have
greater information
precision and increased size of batched orders, and assertions 3
and 4 suggest that
inventory-holding costs should diminish accordingly.
Hypothesis 2: a set of two, autonomous suppliers in a supply
chain, both sharing
demand information from a C3 system, will have decreased

inventory-holding
costs when compared to a supply chain with only the
downstream supplier using
a C3 system.
We expect this outcome because the two suppliers have the
added advantage of
sharing information and gain the advantages of doing so as
suggested in assertion 1.
Hypothesis 3: a set of two, centrally controlled, suppliers in a
supply chain, both
sharing demand information from a C3 system, will have
decreased inventory-
holding costs when compared to a supply chain using
decentralized control.
We expect this outcome because of assertion 2 ‘‘Centralized
control shows greater
overall supply-chain performance than decentralized control and
the performance
differences will increase as delivery windows decrease.’’
4. Research methodology
To evaluate the effects of a C3 system on a supply chain, we
employ a simulation
methodology. The use of simulations in research has broadened
over time to include
a widening range of applications, ranging as far as to evaluate
descriptions of

bounded rational but adaptive economic systems (Cyert and
March 1963, Crecine
1969), normative implications characterized by intentional
rational economic
behaviour (Nelson 1982), as well as issues as complex as
organizational learning
(Levinthal and March 1988, Lant and Mezias 1990). Within the
topic of this paper,
simulations have also been used to evaluate supply-chain
characteristics (Fulkerson
and Staffend 1997, Bhaskaran 1998, Kaihara 2001, Raghunathan
2001, Missbauer
2002, Persson and Hagger 2002, Zhao et al. 2002).
The limitations of simulations include the fact that they are
dependent upon
researchers’ theoretical assumptions and the initializing values
of the independent
variables. Some of these concerns may be obviated by changing
the assumptions of
the model or by altering the initial variables or both, and
performing a sensitivity
analysis over the manifold versions of the model. The problems
of simulations
generally include, as with any abstraction, their omission of
detail, limiting
somewhat the complexity of individuals and decisions and
processes. Additional
problems include attempting to verify the models under
consideration.
To capture the salient characteristics of the supply chain in our
study, we began
by recognizing that there are five elements of a simulation
(Whicker and Seligman
1991).

1. Researcher specified assumptions about the model being
tested.
2. Parameters (fixed values and control variables).
3. Inputs, or independent variables.
4. Algorithms, or process decision rules that convert input
values to outputs.
5. Outputs, or dependent variables.
We explain these elements in the sections that follow, and detail
the simulation
implementation.
4.1 Modeling assumptions
Our supply-chain simulation model is derived from the
modeling elements defined by
Strader et al. (1998, 1999). Previously, we modeled a product
postponement supply
chain (Robinson and Elofson 2000, 2001), a key step in the C3
model, and validated
that simulation model against the analytic model of Anand and
Mendelson (1998).
Herein, we report our analysis of the supply-chain model, which
now includes the use
of the C3 system.
The supply-chain model is abstract. In it, each node can
represent a workstation
within the context of job scheduling in a production facility.
Alternatively, each node
can represent a site within multinational cooperating companies.
Thus, times and
costs are relative within the abstract model.
Figure 3 illustrates the structure of the supply chain. Four kinds

of nodes are
shown: (1) supplier; (2) manufacturer; (3) distributor; and (4)
customer. Together,
manufacturing nodes M2.1 and M2.2 define one aggregate
manufacturing facility –
their intra-communication times for orders and parts are less
than the times for
communicating with the external nodes, M1 and D. Moreover,
manufacturing nodes
M2.1 andM2.2 cooperate to produce timely personalized
products through their
complementary polices of build-to-forecast (BTF) and make-to-
order (MTO),
respectively. Manufacturing node M1 uses a BTF policy.
In figure 4, a broker employing a C3 system (B) is inserted
between the
distributor and the manufacturing nodes for those supply chains
that use the C3
system. The placement of the C3 system node (B) models either
of the real-world
implementations cases in which: (1) the distributor uses the C3
system or (2) the most
downstream manufacturer uses the C3 system.
4.2 Fixed simulation parameters
We defined two supply models that differ only in their use of
the C3 system.
Compare figure 3 with figure 4. Both models have parameters
that were fixed
identically during all simulations. They included the following

parameters:
. Model parameters: in addition to the basic structure of the
supply chain, both
models also define how orders and parts move between and
within nodes of
the supply chain. Numeric values define the order and part
forwarding times,
production times, inventory costs, and start-up inventory
quantities. In the
simulations, the start-up inventory quantity is 20 for each
product type.
Materials & materials information flow
Demand information flow
Z is produced from X and Y: ⇒
S
⇒
⇒
⇒
M1
0,1 3
4
5
0,2
1,2
⇒
⇒
⇒

M2,1
3,4 6
7
8
3,5
4,5
⇒
⇒
⇒
M2,2
6,7 9
10
11
6,8
7,8
⇒
⇒
⇒
D C
9 9
10
11
9
10
11

10
11
X,Y Z
Aggregate node
S
Figure 3. Supply chain without C3 system.
Materials & materials information flow
Demand Information Flow
Z is produced from X and Y: ⇒X,Y Z
Aggregate node
S
M1
0,1 3
0,2 4
1,2 5
⇒
⇒
⇒
⇒
⇒
⇒
⇒
⇒

M2,1
3,4 6
3,5 7
4,5 8
⇒
⇒
⇒
⇒
⇒
⇒
⇒
⇒
M2,2
6,7 9
6,8 10
7,8 11
⇒
⇒
⇒
⇒
⇒
⇒
⇒
⇒
B
9 9
10 10
11 11⇒
⇒
⇒

D C
9 9
10 10
11 11
⇒
⇒
⇒
⇒
⇒
⇒
9
10
11
Figure 4. Supply chain with C3 system, as supported by the
broker B.
Each product production time is 0.01 hours. Inventory costs are
computed as
the holding time multiplied by a fixed cost: (PartexitTime �
PartarrivalTime) �
cost. For all nodes, the cost is $1.00 per part.
. Order and part parameters: all orders originate at customers,
pass through the
distribution node and then onto the manufacturing nodes. The
C3 system
model includes a C3 system between the distribution and
manufacturing
nodes. All parts originate at a supplier and are forwarded to the

manufacturing nodes.
. Part parameters: products are manufactured from parts, as
shown in the
figures. The finished product 9, for example, is manufactured
from parts 6
and 7, which in turn are manufactured from parts 3, 4, and 5,
which are
manufactured from parts 0, 1, and 2. Thus, a single finished
product requires
8 parts, 6 manufacturing processes, 3 manufacturing nodes, as
well as the
supplier and distributor nodes. Thus, small changes in demand
can cascade
into substantial inventory changes. Manufacturing nodes use a
production
plan horizon of 8 hours, which models three production
schedules in
24 hours.
. Demand variability: the market demand for a quantity of a
product at time
point t is represented as normal distribution over a range [low,
high]. At the
time each order enters the simulation, the simulator randomly
selects a
quantity from the distribution. The ranges vary with the demand
pools, as
described in the next section (x4.3).
. Demand forecasting: nodes M1 and M2.1 use a simple
exponential forecasting
function to predict demand quantity for their products. The
formula is as
follows:

Forecasted Demand (t2) ¼ �� Actual Demand (t1) þ (1 � �) �
Forecasted
Demand (t1).
Actual Demand (t1) is the actual demand from previous period.
Forecasted Demand is the forecasted demand from previous
period. �2 [0,1];
it weights the actual vs. forecasted demand.
The forecast horizon is 2 days, and the � parameter is 0.5.
. Simulation time: all simulations run for 45 days with a time
step of 1.0 hour.
4.3 Independent simulation variables
Information sharing, C3 system, and order lead times are
independent simulation
variables.
4.3.1 Information sharing. In the simulation, information
sharing is modeled as the
communication of order demand and inventory levels between
nodes. Three levels of
information sharing are considered:
1. None: no information is shared between nodes.
2. Some: demand information, comprised of product type,
quantity and due
date, is shared among nodes. Demand information flows
upstream, from the
customers toward the suppliers.

3. Most: demand information and supply information, comprised
of inventory
levels, is shared between nodes. Supply information flows
downstream, from
the suppliers toward the customers.
4.3.2 C3 software. In the simulation, the effects of the C3
system are modeled by
two information improvements: (1) increased precision in the
demand information;
and (2) increased demand homogeneity, specifically, orders are
aggregated into pools
having common characteristics. The improvements are modeled
as follows:
. Information precision: generally, one can assume that the
demand informa-
tion (DI) received by a supply-chain node is accurate. That is,
DIreceived ¼ DIactual. However, demand information can be
inaccurate. Let
� be the precision of the demand information (DI). Then, we
model DIreceived
as DIreceived ¼ DIactual þ (DIactual � �) Thus, if � is 0, then
the demand
information is accurate; that is, DIreceived ¼ DIactual.
However, if � is 1, then
the demand received is twice the actual demand.
. Demand pooling: demand pooling aggregates multiple demand
orders
occurring in a period into a single ‘‘batched order,’’ or pool.
Consider the
order arrival rate, which is the demand quantity for a specified
period: arrival
rate ¼ DQ/(t2 t2 � t1). The average arrival rate for some
sequence of periods

i..n, Pi ¼ tiþ1 � ti is DQi/Pi. Demand pooling aggregates the
demand over a
sequence of periods into a single demand order: demand pool ¼
� DQi. Of
course, larger demand pools imply longer demand periods
because pools are
created by aggregating earlier orders with later orders.
In the simulation, two levels of the C3 system are considered:
1. No use of the C3 system.
. Without the C3 system, � ¼ [0.0, 0.1], !�"�"� is drawn from
the normal
distribution over the range. Thus, the non-C3 system model has
up to 10
percent demand imprecision.
Without the C3 system, the demand pools are one-half that of
the C3 system
model. The DQ is drawn the normal distribution over the range
[5, 15], where the
period, Pi, is 24 hours.
2. Use of the C3 system.
. With the C3 system, � ffi, which specifies perfect demand
precision.
. With the C3 system, the DQ is drawn the normal distribution
over the
range [5,15], where the period, Pi, is 48 hours.
The combined modeling elements define the C3 system supply
chain with better
demand information and greater demand pooling than the non-

C3 system supply
chain.
4.3.3 Order lead times. Nine order lead times are analysed. Lead
times ranged from
0 to 32 hours at 4-hour increments (e.g. 0, 4, 8, 16, etc.). The
longest lead time, of
32 hours is four times the production planning period.
Variability is introduced into
lead times by adding a random value �, where � is selected
from the normal
distribution over the range [0.0, 2.0] hours.
4.4 Simulation functions
A simulation begins with the arrival of orders. Each node of the
supply chain fulfils
orders that arrive from its upstream nodes by producing
products from parts that
arrive from its downstream nodes. Two production policies are
used: BTF and
MTO. These basic simulation functions are described next.
. Order arrival: orders arrive for each product after each arrival
interval
elapses. Each order has an associated quantity described by a
normal
distribution. A supply-chain node that receives an order can
send orders to
its own suppliers for any necessary parts. Thus, a customer
order can create a

cascade of orders through the supply-chain network. In the
simulation, a
node fulfils orders according to its production policy under the
constraints of
the fixed simulation parameters (see the section entitled ‘‘Fixed
simulation
parameters’’).
. Make to order production: in a MTO production policy, a node
produces
products to fulfill specific orders. Expedited orders are allowed.
Thus, late
arriving orders with short lead times, may be inserted in front of
existing
orders in the production plan.
. Build to forecast production: in a BTF production policy, a
node produces
products as a means to fulfill an expected quantity of future
orders, as
determined by a forecast (Mccutcheon et al. 1994). As orders
arrive at a
node, the node can continually update its expectation of future
orders. (The
section entitled ‘‘Fixed simulation parameters’’ defines the
exponential
forecasting function.)
. Information sharing: supply-chain nodes can share
information. For example,
an upstream node can share its demand forecast with a
downstream node.
This allows the downstream node to update its demand forecast
to be
consistent with the upstream node. Figures 3 and 4 illustrate
such

information sharing.
. Part arrival: part arrival is analogous to order arrival. Parts
arrive at supply
nodes after each arrival interval elapses. Each part shipment has
an
associated quantity described by a normal distribution.
4.5 Dependent simulation variables
Each simulation run tracks a number of dependent variables,
including: cycle time,
fulfillment rate, inventory cost, and capital utilization.
Inventory cost is the focus on
the following experiments and discussion. However, cycle time
is also referenced.
Their simulation functions are defined here.
. Inventory cost: inventory costs are computed for each node. It
is the
holding time multiplied by a fixed cost: (PartexitTime �
PartarrivalTime) � cost.
The analysis in the following sections on Experiments and
Discussion refer to
the total inventory costs for all nodes of the supply chain
excluding the
supplier, distributor, and customer.
. Cycle time: cycle times are computed for the distributor node
of each supply
chain. It is the order fulfillment period: (PartexitTime �
OrderarrivalTime).

These dependent variables are used to assess the effects of the
experiments.
4.6 Simulation implementation
Our simulations are built using simulations tools. The supply-
chain design uses some
components defined by the Strader et al. simulation framework
(SSM) (Strader et al.
1998, 1999). A simulation framework is a carefully designed set
of reusable
components that can be specialized to produce custom
simulations. The Strader
et al. SSM defines configurable supply-chain components for
order management,
inventory management, production planning, capacity planning,
materials planning,
shop floor control, and manufacturing systems. SSM is built
upon Swarm, a general-
purpose agent-based simulation framework.
Swarm is a multi-agent software platform for the simulation of
complex adaptive
systems. In the Swarm system the basic unit of simulation is the
swarm, a collection
of agents executing a schedule of actions. Swarm supports
hierarchical modeling
approaches whereby agents can be composed of swarms of other
agents in nested
structures (Minar et al. 1996).
Our simulation implementations benefit from the pre-defined
supply-chain
components of SSM, and the general simulation processing and
programming

environment of Swarm.
The top of figure 5 illustrates the SSM simulation network
defined in Strader
et al. (1998), where each entity node is defined as interacting
agents, as shown at the
bottom of the figure. SSM configuration parameters include the
supply-chain
network design and the inputs, outputs, and algorithms of the
agents. Configuration
is specified as parameters in SSM, as well as data and programs
in the underlying
Swarm framework.
In Strader et al. (1999), Strader et al. describe the interactions
among the agents.
[. . .] an entity ScnESwarm A receives an order from its
customer ScnESwarm C.
The order flows to the order management agent (OrdM).
According to the
customer lead times, the inventory availability information
(from InvM),
the production plan (from PrdP), and the manufacturing
capacity (CapP), the
order management agent assigns a due date to the order. If the
products are
in stock, the order is filled by shipping the products from
inventory. If the
products are in receiving, the due date is set according to the
delivery date of the
products.
For an entity with manufacturing capability, the order is
forwarded to the
production-planning agent (PrdP) where the schedule for

making the products is
planned. The capacity-planning agent (CapP) and the material-
planning agent
(MatP) are partner agents in generating achievable build plans.
The material
planning obtains build plans from the production-planning agent
to allocate
materials for manufacturing. It also contributes information
about material
availability to production planning for scheduling. The capacity
planning agent
(CapP) plans capacity by taking the build plan from PrdP and
sends capacity
usage information to PrdP for scheduling the build plan. The
SCN management
agent (ScnM) takes the order information to choose suppliers in
allocating
material sources
As can be inferred from the above descriptions, the SSM
simulation framework
provides a rich, detailed, accurate representation of a supply-
chain network.
Our use of the SSM follows the paradigm of software product
line engineering,
which addresses the systematic reuse of core assets for building
related products

(Clements and Northrop 2002, Pohl et al. 2005). A product line
is a set of similar
products that are derived from common components and yet
possess varied features
(a)
(b)
Figure 5. An example supply chain (top), where each node is
defined as multiple interacting
agents (bottom). Figures adapted from Strader et al. (1998).
to meet specific customer requirements. Each individual product
within a product
line is called a product variant. Our simulations are product
variants derived from the
SSM framework. As a consequence of using SSM, we gain the
benefit of individually
validated components that can be adapted and integrated to
create new supply-chain
designs.
Our simulation designs vary from the SSM defined in Strader et
al. (1998), as
depicted at the top of figure 5. For example, our designs specify
6- and 7-tier
networks, as figures 3 and 4 show, in contrast to the 5-tier
model defined in Strader
et al. (1998). Other variations include different kinds of

components (e.g. multiple
manufacturing nodes, a newly defined broker node), the
simultaneous use of multiple
policies (MTO, BTF), as well all of the simulation parameters
(see the section
entitled ‘‘Fixed simulation parameters’’).
4.7 Benchmarking the models
Law and Kelton suggest six techniques for increasing the
validity and credibility of a
simulation model (Law and Kelton 2000). These include (1)
conversations with
subject-matter experts; (2) observations of this system; (3)
comparison with existing
theory; (4) comparison with similar simulation studies; (5)
experience and intuition
of the modelers; and (6) animation. We have applied all of these
techniques to our
simulations. Techniques 3 and 4 are perhaps the most important
because of the
precision and strength of their argumentation.
The introductory and concluding sections draw comparison with
existing theory.
We compared our simulations with prior studies, including
analytic studies (Anand
and Mendelson 1998), our own prior simulations (Robinson and
Elofson 2000,
2001), and results reported in Strader et al. (1998, 1999). Our
simulated inventory
costs are consistent with Anand and Mendelson’s analytically
derived costs (Anand
and Mendelson 1998), and our manufacturing policies are
consistent with Strader
et al. (1998, 1999): that is, we use the same policies but do not

simulate the same
events.
5. Experiments
Here, we report the results of 63 simulations. To provide an
understanding of how
these simulations were generated, table 1 is presented. It can be
used to compare the
results relating the C3 system, information sharing, and lead
time.
Table 1. An empty result form for six simulations with a lead
time of 12 hours.
Lead time ¼ 12 hours C3 system Non-C3 system
No sharing
Some sharing
Most sharing
Table 1 illustrates an empty experiment report table for
comparing the C3 system
and the non-C3 system for three levels of information sharing,
with a lead time of
12 hours. (The dependent variable values are not shown.) This
table illustrates just
one of a number of experiments run. We report on 63 simulation
runs (2 C3 system �
3 sharing � 9 lead times).
Figure 6 shows the results of the experiments. Lead times are

plotted along the
x-axis, and the total supply-chain inventory costs are plotted
along the y-axis. Each
plotted point represents an experiment drawn from the
combinations of C3
system � Sharing � Lead times. The lowest three lines plot
three sharing levels over
the nine lead times for a supply chain using the C3 system,
while the upper three lines
show the same for a non-C3 system supply chain.
Figure 6 indicates that use of a C3 system reduces inventories
substantially, and
sharing reduces inventories, especially when the C3 system is
unavailable. Lead times
have little effect on inventories, except where lead time is 0 þ
[0.0, 2.0].
A two-tailed paired t-test comparison between the C3 system
and non-C3 system
for no sharing, some sharing, and most sharing indicates p
values of 7.026 � 10
�7
,
1.70 � 10
�8
, and 1.97 � 10
�8
, thus supporting the hypotheses that the C3 system
makes a difference in the three levels of sharing across the lead
times.

6. Discussion
Earlier, we posed three hypotheses to test. Here, we discuss how
the simulation
results illuminate the hypotheses.
$-
$100,000
$200,000
$300,000
$400,000
$500,000
$600,000
$700,000
0 4 8 12 16 20 24 28 32
Lead time (hr)
In
v
e
n
to
ri
e

s
C3, sharing most C3, sharing some C3, no sharing
no C3, sharing most no C3, sharing some no C3, no sharing
Figure 6. Inventory costs for the combinations of C3 system,
information sharing,
and lead times.
Table 2 summarizes the hypotheses analysis, which are
computed from the
simulation results summarized in the appendix. Each hypothesis
is confirmed.
6.1 H1: using C3 system will decrease inventory holding costs
Hypothesis H1 states that:
H1: a supplier using a C3 system will have decreased inventory
holding costs
compared to a supplier with no C3 system.
Table 3 reveals that the downstream C3 system-assisted node in
the supply chain
(C3 system M2) had lower inventory holding costs over the
corresponding unassisted
node (No C3 system M2). Over the full range of lead times,
savings were incurred in
keeping with assertion 3: a C3 system batching capability will
likely lower inventory
holding costs, and assertion 4: a C3 system’s increased
information precision will likely

lower inventory holding costs. Thus, the addition of a C3
system to the downstream
node in our supply chain resulted in savings for that node.
Based on the work of Anand and Mendelson (1998), Kulp
(2002) and Moinzadeh
(2002), this is because the C3 system-assisted node has greater
information precision
in its orders as well as batched orders. Table 3 shows the
comparison of the two
approaches.
Here, we note that the average inventory savings across all lead
times (with no
sharing) for M2 was $239,462. The percent savings over these
averages was 79.1%.
For the entire supply chain, we averaged the total inventory
costs for both nodes M1
and M2, comparing the non-C3 supply chain with the C3 supply
chain. Here, the
overall supply chain derived savings of $602,046 using a C3
system, for an 89.3%
savings. Moreover, these results indicate that the supply chain,
on the whole,
experienced savings in keeping with that reported in earlier
work.
Table 2. Summary of the hypotheses and simulation results.
Hypothesis Simulation results
H1: a supplier using C3 system
will have decreased inventory-holding
costs compared to a supplier
with no C3 system

Confirmed: average savings of 89.3%
H2: a set of two, autonomous suppliers
in a supply chain, both sharing demand
information from a C3 system,
will have decreased inventory-holding
costs when compared to a supply chain
with only the downstream supplier
using a C3 system
H3: a set of two, centrally controlled,
suppliers in a supply chain,
both sharing demand information
from a C3 system, will have decreased
inventory-holding costs when compared
to a supply chain using decentralized control
T
a
b
le
3
.
A
v
er

a
g
e
co
st
s
a
n
d
th
e
p
er
ce
n
ta
g
e
im
p
ro
v
em
en
t.
C
3
sy
st
em
N
o

C
3
sy
st
em
D
if
fe
re
n
ce
P
er
ce
n
ta
g
e
sh
a
ri
n
g
M
1
M
2
T
o
ta

l
M
1
M
2
T
o
ta
l
M
1
M
2
T
o
ta
l
M
1
M
2
T
o
ta
l
n

o
n
e
$
9
,0
0
2
$
6
3
,3
1
2
$
7
2
,3
1
4
$
3
7
1
,5
8
6
$
3
0
2

,7
7
4
$
6
7
4
,3
6
0
$
3
6
2
,5
8
4
$
2
3
9
,4
6
2
$
6
0
2
,0
4
6

9
8
%
7
9
.1
%
8
9
.3
%
so
m
e
$
3
,8
2
8
$
2
7
,4
5
3
$
3
1
,2
8

1
$
1
3
7
,1
7
7
$
1
2
4
,1
9
2
$
2
6
1
,3
6
9
$
1
3
3
,3
4
9
$
9

6
,7
3
9
$
2
3
0
,0
8
9
9
7
%
7
7
.9
%
8
8
.0
%
m
o
st
$
3
,7
6
7

$
2
2
,7
3
6
$
2
6
,5
0
2
$
9
3
,7
1
6
$
9
0
,9
4
9
$
1
8
4
,6
6
5

$
8
9
,9
4
9
$
6
8
,2
1
3
$
1
5
8
,1
6
2
9
6
%
7
5
.0
%
8
5
.6
%

6.2 H2: sharing C3 system data will decrease inventory holding
costs
H2: a set of two, autonomous suppliers in a supply chain, both
sharing demand
information from a C3 system, will have decreased inventory
holding costs when
compared to a supply chain with only the downstream supplier
using a C3 system.
Figure 6 illustrates that a supply chain with a C3 system and
sharing demand
information (second from bottom line in figure 6) does reduce
overall inventory
holding costs when compared to supply chain with just C3 and
no demand sharing
(third from bottom line in figure 6). For the full set of lead
times, savings were
incurred in keeping with assertion 1: as information sharing
increases across a supply
chain, inventory savings increase, as well as assertions 3 and 4.
Considering the research of Chen (1998), Cachon and Fisher
(2000), Lee et al.
(2000) and Moinzadeh (2002), we expected savings up to 34.9%
even without sharing
C3 system information. In our simulation, providing upstream
suppliers with C3
system data results in the demand information’s precision being

improved and not
surprisingly, the savings improving as well.
Table 4 summarizes the findings for this simulation. The supply
chain using the
C3 system with some shared information had an average
inventory cost reduction of
$41,033. The percent savings of the C3 system version over the
non-C3 system was
56.7%. Also, when we consider the inventory savings of the
downstream node when
sharing versus not-sharing, we see that it is in the M2 node’s
interest to share demand
information with its M1 supplier (saving $35,860 over non-
sharing supplier).
6.3 H3: sharing C3 system data and inventory data will decrease
inventory holding
costs more than sharing C3 system data alone
H3: a set of two, centrally controlled, suppliers in a supply
chain, both sharing
demand information from a C3 system, will have decreased
inventory holding
costs when compared to a supply chain using decentralized
control.
Table 4. C3 system emphasizing some information sharing.
C3 system
sharing M1 M2 Total
none $9,002 $63,312 $72,314

some $3,828 $27,453 $31,281
most $3,767 $22,736 $26,502
Difference
none-some $5,174 $35,860 $41,033
some-most $61 $4,717 $4,778
none-most $5,235 $40,577 $45,812
Percentage
none-some 57.5% 56.6% 56.7%
some-most 1.6% 17.2% 15.3%
none-most 58.2% 64.1% 63.4%
Figure 6 illustrates that a supply chain with a C3 system and
full sharing – demand
information and inventory information – (bottom line in figure
6) does reduce
overall inventory holding costs when compared to supply chain
with C3 and demand
sharing (second from bottom line in figure 6). For the full set of
lead times, savings
were incurred in keeping with assertion 2: centralized control
shows greater overall
supply-chain performance than decentralized control, and the
performance differences
will increase as delivery windows decrease.
However, while the results of the simulation find in favour of
the hypothesis, we

did not observe greater performance differences as the delivery
window decreased.
This may be because our simulation did not allow for the
possibility of failing to fill
an order because of delivery times being less than final stage
MTO process (often a
prelude to switching to a build-to-stock policy (Watson and
Polito)).
Table 5 summarizes the findings for this simulation. The supply
chain using the
C3 system with only some shared information had an average
inventory cost of
$41,033, while the supply chain sharing full information had an
average inventory
cost of $45,812. The percent savings of the full information
sharing version over the
C3 system with some information sharing was 15.3%. These
results suggest that
centralized control, when exercising the prerogative of sharing
full information,
benefits more from doing so than not.
7. Conclusions
The limitations of this study are consistent with and
characteristic of those found in

simulation studies in general: the researchers’ theoretical
assumptions as well as the
initializing values of the independent variables influence the
outcomes of the
experiments. To obviate this as much as possible, the
simulations were benchmarked
against existing analytical models and used previously validated
simulation models.
Notwithstanding these efforts, as with any abstraction, there is
an omission of detail
that would be encountered in an in situ empirical study.
Table 5. C3 system emphasizing most information sharing.
C3 system
sharing M1 M2 Total
none $9,002 $63,312 $72,314
some $3,828 $27,453 $31,281
most $3,767 $22,736 $26,502
Difference
none-some $5,174 $35,860 $41,033
some-most $61 $4,717 $4,778
none-most $5,235 $40,577 $45,812
Percentage
none-some 57.5% 56.6% 56.7%
some-most 1.6% 17.2% 15.3%
none-most 58.2% 64.1% 63.4%

Overall, the simulations supported hypotheses 1, 2, and 3. The
assertions
regarding information precision, batch orders, information
sharing, and centralized
control offer, we believe, plausible rationales for these
outcomes, especially given the
body of research that formed the foundation of these assertions.
Further, these
outcomes are consistent with the theoretical predictions posited
by Berger et al.
(2005).
Thus, a manufacturer employing a C3 system might reasonably
expect to find
inventory savings with its use. That same manufacturer might,
with conditions
permitting, also find even greater savings if it shares the C3
system data with its
upstream supplier, and more so if it shares its inventory
information with its
upstream supplier. With this information, managers may be able
to make sourcing
decisions that best capitalize on their alternative, choosing

those partners that are,
ceteris parabis, most suitable for sharing customer order and
inventory information
(given conditions such as regulations, etc., as mentioned earlier
(Lee and Whang
2000)).
The implications here are that, while the initial costs of a C3
system may be
greater than traditional market research, the downstream
savings may offset these
costs. Further, desirable side effects might include: (1)
increased customer switching
costs, because of the knowledge and time the customers devote
to the development of
their product; and (2) improved production planning because of
early customer
commitment to purchase.
Beyond our hypotheses, it should be noted that the C3 system-
facilitated supply
chains in our study always showed inventory savings over their
non-C3 system
comparables. Also, the data suggests that sharing demand
information had a
generally greater positive affect on inventory savings than
sharing inventory

information, although both were positive even when taken
together.
A C3 system approach is not suitable for all products. A greater
understanding of
the parameters that govern the choice of using C3 should
provide a rich opportunity
for research. The many perspectives that may illuminate these
questions range from
relationship marketing, to human factors, to studies of trust, to
group systems, as
well as additional production questions. For example, are there
design processes that
have intermediate results which would allow for postponement
strategies to be better
employed? Also, can an automated interaction be created that
gives customers
feedback on the costs of their design while they are being
investigated, creating cost
scenarios with a what-if analysis as they work? Future research
might also consider
the costs of introducing and maintaining a C3 system. Then, a
comparison of the C3
system costs against the savings obtained by the C3 system
would further specify the

circumstances under which a C3 system will be advantageous.
In conclusion, we have presented an IT system that uses
customer knowledge and
information to collaboratively create product designs, and
discussed how this
knowledge is managed forward in the production planning
process. We have tested
how the implementation of this IT system affects inventory
saving via information
sharing, large production runs, and more to help manufacturers
think about supplier
selection and partnership formation in supply chains. Finally,
we illustrated how this
knowledge sharing can be useful to the firm.
A
p
p
e
n
d
ix
A
n
a

ly
si
s
ta
b
le
C
3
sy
st
em
N
o
C
3
sy
st
em
D
if
fe
re
n
ce
P
er
ce
n
ta
g
e

S
h
a
ri
n
g
N
1
N
2
T
o
ta
l
N
1
N
2
T
o
ta
l
N
1
N
2
T

o
ta
l
N
1
N
2
T
o
ta
l
N
o
n
e
$
9
,0
0
2
$
6
3
,3
1
2
$
7
2

,3
1
4
$
3
7
1
,5
8
6
$
3
0
2
,7
7
4
$
6
7
4
,3
6
0
$
3
6
2
,5
8
4

$
2
3
9
,4
6
2
$
6
0
2
,0
4
6
9
8
%
7
9
.1
%
8
9
.3
%
S
o
m
e
$

3
,8
2
8
$
2
7
,4
5
3
$
3
1
,2
8
1
$
1
3
7
,1
7
7
$
1
2
4
,1
9
2
$

2
6
1
,3
6
9
$
1
3
3
,3
4
9
$
9
6
,7
3
9
$
2
3
0
,0
8
9
9
7
%
7
7

.9
%
8
8
.0
%
M
o
st
$
3
,7
6
7
$
2
2
,7
3
6
$
2
6
,5
0
2
$
9
3
,7

1
6
$
9
0
,9
4
9
$
1
8
4
,6
6
5
$
8
9
,9
4
9
$
6
8
,2
1
3
$
1
5
8

,1
6
2
9
6
%
7
5
.0
%
8
5
.6
%
D
if
fe
re
n
ce
n
o
n
e-
so
m
e
5
,1
7

4
3
5
,8
6
0
4
1
,0
3
3
2
3
4
,4
0
8
1
7
8
,5
8
2
4
1
2
,9
9
1
2

2
9
,2
3
4
1
4
2
,7
2
3
3
7
1
,9
5
7
S
o
m
e-
m
o
st
6
1
4
,7
1
7

4
,7
7
8
4
3
,4
6
2
3
3
,2
4
3
7
6
,7
0
5
4
3
,4
0
0
2
8
,5
2
6
7

1
,9
2
6
n
o
n
e-
m
o
st
5
,2
3
5
4
0
,5
7
7
4
5
,8
1
2
2
7
7
,8
7
0

2
1
1
,8
2
5
4
8
9
,6
9
5
2
7
2
,6
3
5
1
7
1
,2
4
9
4
4
3
,8
8
3

P
er
ce
n
ta
g
e
n
o
n
e-
so
m
e
5
7
.5
%
5
6
.6
%
5
6
.7
%
6
3
.1
%

5
9
.0
%
6
1
.2
%
6
3
.2
%
5
9
.6
%
6
1
.8
%
so
m
e-
m
o
st
1
.6
%

1
7
.2
%
1
5
.3
%
3
1
.7
%
2
6
.8
%
2
9
.3
%
3
2
.5
%
2
9
.5
%
3

1
.3
%
n
o
n
e-
m
o
st
5
8
.2
%
6
4
.1
%
6
3
.4
%
7
4
.8
%
7
0
.0
%

7
2
.6
%
7
5
.2
%
7
1
.5
%
7
3
.7
%
References
Anand, K.S. and Mendelson, H., Information and organization
for horizontal multi-market
control. Manage. Sci., 1997, 43, 1609–1627.
Anand, K.S. and Mendelson, H., Postponement and information
in a supply chain. Discussion
Papers (1222). Northwestern University, Center for
Mathematical Studies in
Economics and Management Science, 1998.

Aviv, Y., The effect of collaborative forecasting on supply
chain performance. Manage. Sci.,
2001, 47, 1326–1343.
Berger, C., Möslein, K., Piller, F. and Reichwald, R., Co-
designing modes of cooperation
at the customer interface: learning from exploratory research.
Eur. Manage. Rev., 2005,
2, 70.
Bhaskaran, S., Simulation analysis of manufacturing supply
chain. Decision. Sci., 1998, 29,
633–657.
Boynton, A.C. and Victor, B., Beyond Flexibility: building and
managing the dynamically
stable organization. Calif. Manage. Rev., 1991, 3, 53–66.
Cachon, G., Managing supply chain demand variability with
scheduled ordering policies.
Manage. Sci., 1999, 45, 843–856.
Cachon, G.P. and Fisher, M., Supply chain management and the
value of shared information.
Manage. Sci., 2000, 46, 1032–1048.
Chen, E., Echelon reorder points, installation reorder points,
and the value of centralized
demand information. Manage. Sci., 1998, 44, 221–234.
Chen, F.G., Decentralized supply chains subject to information
delays. Manage. Sci., 1999,
45, 1076–1090.
Cheung, K.L. and Lee, H.L., The inventory benefit of shipment

coordination and stock
rebalancing in a supply chain. Manage. Sci., 2002, 48, 300–306.
Clements, P. and Northrop, L., Software Product Lines:
Practices and Patterns, 2002
(Addison-Wesley: Boston, MA).
Crecine, J.P., A Computer Simulation Model of Municipal
Budgeting, 1969 (Rand McNally:
Chicago, IL).
Cvsa, V. and Gilbert, S.M., Strategic commitment versus
postponement in a two-tier supply
chain. Eur. J. Oper. Res., 2002, 141, 526.
Cyert, R.M. and March, J.G., A Behavioral Theory of the Firm,
1963 (Prentice Hall:
Englewood Cliff, NJ).
Dang, J. and Huhns, M.N., Coalition Deal Negotiation for
Services. , In Proceedings of the
First International Workshop on Rational, Robust, and Secure
Negotiations in
Multi-Agent Systems (RRS’05), edited by Ito, T., Hattori, H.,
Matsuo, T. and Zhang,
M., 2005, edited by Ito, T., Hattori, H., Matsuo, T. and Zhang,
M., (IEEE Computer
Society Press).
De Meyer, A., Dutta, S. and Srivastava, S., The Bright Stuff:
How Innovative People and
Technology Can Make the Old Economy New, 2002 (Prentice-
Hall: London).
Elofson, G.S. and Robinson, W.N., Creating a custom mass-
production channel on the

Internet. Commun. ACM, 1998, 41, 56–62.
Fulkerson, B. and Staffend, G., Decentralized control in the
customer focused enterprise.
Ann. Oper. Res., 1997, 77, 325–333.
Gascó-Hernández, M. and Torres-Coronas, T., Virtual teams and
their search for creativity.
In Virtual and Collaborative Teams, edited by S.H. Godar and
S.P. Ferris, 2004
(Idea Group: Hershey, PA).
Gibbert, M., Leibold, M. and Probst, G., Five styles of customer
knowledge management,
and how smart companies use them to create value. Eur.
Manage. J., 2002, 20,
459–469.
Heikkila, J., From supply to demand chain management:
efficiency and customer satisfaction.
J. Oper. Manage., 2002, 20, 747.
Herlocker, J.L., Konstan, J.A., Terveen, L.G. and Riedl, J.T.,
Evaluating collaborative
filtering recommender systems. ACM Trans. Inf. Syst., 2004,
22, 5–53.
Homburg, C., Workman, J.P.J. and Jensen, O., Fundamental
changes in marketing
organization: the movement toward a customer-focused
organizational structure.
Acad. Mark. Sci. J., 2000, 28, 459.

Hyung Rim, C., Hyun Soo, K., Soon Goo, H., Young Jae, P.,
Young Sung, P. and Moo
Hong, K., Implementation of Framework for Developing Multi-
Agent Based
Automated Negotiation Systems. In Proceedings of the 7th
international conference on
Electronic commerce, 2005, ACM Press, Xi’an, China, 2005.
Jin Baek, K. and Segev, A., A framework for dynamic
eBusiness negotiation processes.
IEEE International Conference on E-Commerce, IEEE, 2003.
Kaihara, T., Supply chain management with market economics.
Int. J. Prod. Econ., 2001, 73,
5–14.
Khalid, H.M. and Helander, M.G., Web-based do-it-yourself
product design. In The
Customer Centric Enterprise: Advances in Mass Customization
and Personalization,
edited by M. Tseng and F. Piller, 2003 (Springer: New York,
Berlin).
Kouvelis, P. and Gutierrez, G.J., The newsvendor problem in a
global market: optimal
centralized and decentralized control policies for a two-market
stochastic inventory
system. Manage. Sci., 1997, 43, 571–585.
Kulp, S., The effect of information precision and information
reliability on manufacturer-
retailer relationships. Account. Rev., 2002, 77, 653–677.
Lant, T.K. and Mezias, S.J., Managing discontinuous change: a
simulation study of

organizational learning and entrepreneurship. Strategic Manage.
J., 1990, 11, 147–179.
Law, A.M. and Kelton, W.D., Simulation Modeling and
Analysis, 2000 (McGraw Hill:
New York).
Lee, H. and Billington, C., Material management in
decentralized supply chains. Manage. Sci.,
1993, 42, 835–847.
Lee, H., So, K. and Cs, T., The value of information sharing in
a two-level supply chain.
Manage. Sci., 2000, 46, 626–643.
Lee, H. and Whang, S., Decentralized multi-echelon supply
chains: incentives and
information. Manage. Sci., 1999, 45, 633–640.
Lee, H.L., Effective inventory and service management through
product and process redesign.
Oper. Res., 1996, 44, 151–159.
Lee, H.L. and Whang, S., Information sharing in a supply chain.
Int. J. Manuf. Tech.
Manage., 2000, 1, 79.
Levinthal, D.A. and March, J.G., A model of adaptive
organizational search. In Decisions and
Organizations, edited by J.G. March, 1988 (Basil Blackwell:
New York).
Mccutcheon, D.M., Amitabh, S. and Meredith, J.R., The
customization-responsiveness
squeeze. Sloan. Manage. Rev., 1994, Winter, 89–99.

Milgrom, P. and Roberts, J., The economics of modern
manufacturing technology, strategy,
and organization. Am. Econ. Rev., 1990, 80, 511–528.
Minar, N., Burkhart, R., Langton, C. and Askenazi., M., The
swarm simulation system:
a toolkit for building multi-agent simulations, Santa Fe
Institute, 1996.
Missbauer, H., Aggregate order release planning for time
varying demand. Int. J. Prod. Res.,
2002, 40, 699–718.
Moinzadeh, K., A multi-echelon inventory system with
information exchange. Manage. Sci.,
2002, 48, 414–426.
Nelson, R.R. and Winter, S.G., An Evolutionary Theory of
Economic Change, 1982 (Belknap
Press: Boston, MA).
Nemiro, J.E., Connection in creative virtual teams. J. Behav.
App. Manage., 2001, 3, 92–112.
Ogawa, S. and Piller, F., Reducing the risks of new product
development. Sloan. Manage.
Rev., 2006, 47, 65–72.
Persson, F. and Hagger, J., Performance simulations of supply
chain designs. Int. J. Prod.
Res., 2002, 77, 231–245.
Pillar, F.T., Moeslein, K. and Stotko, C.M., Does mass-
customization pay? An economic
approach to evaluate customer integration. Prod. Plan. Cont.,
2004, 15, 435–444.

Piller, F., Schubert, P., Koch, M. and Möslein, K., Overcoming
mass confusion:
collaborative customer co-design in online communities. J.
Comput. Med. Commun.,
2005, 10, article 8.
Piller, F.T., Moeslein, K. and Stotko, C.M., Does mass-
customization pay? An economic
approach to evaluate customer integration. Prod. Plan. Control.,
2004, 15, 435–444.
Pine II, B.J. II, Peppers, D. and Rogers, M., Do you want to
keep your customers forever?
Harvard Bus. Rev., 1995, 73, 103–114.
Pine, J.B., Mass customization: The new frontier in business
competition, 1993 (Harvard
Business School Press: Cambride Mass).
Pohl, K., Böckle, G. and Linden, F.J.V.D., Software Product
Line Engineering: Foundations,
Principles and Techniques, 2005 (Springer: Berlin).
Prahalad, C.K. and Ramaswamy, V., The Future of Competition:
Co-Creating Unique Value
with Customers, 2004 (Harvard Business School Press: Boston,
MA).
Qi-Ming, H., Jewkes, E.M. and Buzacott, J., The value of
information used in inventory
control of a make-to-order inventory-production system. IIE

Trans., 2002, 34, 999.
Raghunathan, S., Information sharing in a supply chain: a note
on its value when demand is
nonstationary. Manage. Sci., 2001, 47, 605–610.
Ramirez, R., Value co-production: intellectual origins and
implications for practice and
research. Strategic. Manage. J., 1999, 20, 49.
Resnick, P. and Varian, H.R., Special issue on recommender
systems. Commun. ACM, 1997,
40.
Robinson, W.N. and Elofson, G., Electronic Broker Impacts on
the Value of Postponement,
in Hawaii. Proceedings of the 33rd Annual Hawaii International
Conference on System
Sciences, IEEE, 2000.
Robinson, W.N. and Elofson, G., Electronic broker impacts on
the value of postponement in
a global supply chain. J. Global Inform. Manage., 2001, 9, 29–
43.
Robinson, W.N. and Volkov, S., Supporting the negotiation life-
cycle. ACM, Commun. ACM,
1998, 41, 95–102.
Rowley, J., Eight questions for customer knowledge
management in e-business. J. Know.
Manage., 2002, 6, 500–511.
Salvador, F., Rungtusanatham, M. and Forza, C., Supply chain
configurations for mass
customization. Prod. Plan. Cont., 2002, 15, 381–397.

Sawhney, M. and Prandelli, E., Communities of creation:
managing distributed innovation in
turbulent markets. Calif. Manage. Rev., 2000, 42, 24.
Schubert, P., The participatory electronic product catalog:
supporting customer collaboration
in e-commerce applications. Electron. Mark. J., 2000, 10, 229–
236.
Seybold, P.B., Marshak, R. and Lewis, J., The Customer
Revolution: How to Thrive When
Customers Are in Control, 2001 (Crown Business: New York).
Sheth, J.N., Sisodia, R.S. and Sharma, A., The antecedents and
consequences of customer-
centric marketing. Acad. Mark. Sci. J., 2000, 28, 55.
Strader, T.J., Lin, F.-R. and Shaw, M.J., Simulation of order
fulfilment in divergent assembly
supply chains. J. Artif. Soc. Soc. Simul., 1998, 1.
Strader, T.J., Lin, F.-R. and Shaw, M.J., The impact of
information sharing on order fulfilment
in divergent differentiation supply chain. J. Global Inform.
Manag., 1999, 7, 16–25.
Su, S.Y.W., Huang, C., Hammer, J., Huang, Y., Li, H., Wang,
L., Liu, Y.,
Pluempitiwiriyawej, C., Lee, M. and Lam, H., An Internet-based
negotiation server
for e-commerce. VLDB J., 2001, 10, 72–90.
Thomke, S. and von Hippel, E., Customers as innovators: a new
way to create value. Harvard
Bus. Rev., 2002, 80, 74–81.

Thonemann, U.W., Improving supply-chain performance by
sharing advance demand
information. Eur. J. Oper. Res., 2002, 142, 81–107.
von Hippel, E., Economics of product development by users: the
impact of ‘‘sticky’’ local
information. Manage. Sci., 1998, 44, 629.
von Hippel, E. and Tyre, M.J., How learning is done: problem
identification in novel process
equipment. Res. Pol., 1995, 24, 1.
Watson, K. and and Polito, T., Comparison of DRP and TOC
financial performance within a
multi-product, multi-echelon physical distribution environment.
Int. J. Prod. Res., 1995,
41, 741–765.
Westbrook, R. and Williamson, P.J., Mass customization:
Japan’s new frontier. Eur. Manage.
J., 1995, 11, 38–45.
Whicker, M.L. and Seligman, L., Computer Simulation
Application: An Introduction, 1991
(Sage Press: Newberry Park, CA).
Wikström, S., Value creation by company-consumer interaction.
J. Mark. Manage., 1996, 12,
359–374.
Zhao, X., Jx, X. and Leung, J., The impact of forecasting model

selection on the value of
information sharing in a supply chain. Eur. J. Oper. Res., 2002,
142, 321–344.
J Intell Manuf (2012) 23:1083–1101
DOI 10.1007/s10845-010-0430-3
Impact of information sharing in hierarchical decision-making
framework in manufacturing supply chains
Nurcin Celik · Sai Srinivas Nageshwaraniyer ·
Young-Jun Son
Received: 14 November 2009 / Accepted: 25 June 2010 /
Published online: 22 July 2010
© Springer Science+Business Media, LLC 2010
Abstract This paper presents a comprehensive framework
for the analysis of the impact of information sharing in hier-
archical decision-making in manufacturing supply chains.
In this framework, the process plan selection and real-time
resource allocation problems are formulated as hierarchical
optimization problems, where problems at each level in the
hierarchy are solved by separate multi-objective genetic algo-
rithms. The considered multi-objective genetic algorithms
generate near optimal solutions for NP-hard problems with
less computational complexity. In this work, a four-level hier-
archical decision structure is considered, where the decision
levels are defined as enterprise level, shop level, cell level,

and equipment level. Using this framework, the sources of
information affecting the achievement of best possible deci-
sions are then identified at each of these levels, and the extent
of their effects from sharing them are analyzed in terms of the
axis, degree and the content of information. The generality
and validity of the proposed approach have been successfully
tested for diverse manufacturing systems generated from a
designed experiment.
Keywords Multi-objective optimization · Supply chain ·
Information sharing · Shop floor control ·
Hierarchical decision-making
Introduction
In today’s global and competitive market circumstances, hav-
ing an efficient supply chain has become more significant
N. Celik · S. S. Nageshwaraniyer · Y.-J. Son (B)
Systems and Industrial Engineering, The University of Arizona,
Tucson, AZ 85721, USA
e-mail: [email protected]
than ever. Many industries in various sectors have launched
initiatives to enhance the efficacy of their supply chain man-
agement, and massive amounts of savings were achieved
as a result of such pioneering works. For example, Kurt
Salmon Associates Inc. (1993) has estimated that potential
savings can add up to around $30 billion per year for the gro-
cery industry. For the healthcare industry, Computer Science
Corp. (1996) estimated that savings would be about $11 bil-
lion per year (48% of the process costs) when the informa-
tion is helped flow in the right direction at right time in the
supply chain. Later, according to Premier Alliance (2004),
“The Supply Chain Collaborative Breakthrough Series”
initiative, which is established by members of Premier Inc.
(an alliance of 1,500 hospitals and health systems), saved

more than $46 million over a 3 year period by enabling the
sharing of the data among various echeclons of the sup-
ply chains. Procter & Gamble (manufacturer) and Wal-mart
(retailer) have disclosed their new supply chain integration
initiative with an aggressive goal of sharing information via
their channel partnership. The partnership started with a sim-
ple desire to improve business relationships, and was gradu-
ally enhanced by sharing information and knowledge about
their respective markets. This sharing in turn enabled more
effective execution of such concepts as category manage-
ment, continuous replenishment, and process coordination,
which collectively helped make the supply chain more effi-
cient (Grean and Shaw 2002; Siems 2005). The mutual bene-
fits that Procter & Gamble and Wal-mart obtained out of this
initiative reached more than billion dollars.
While many other instances of such initiatives as the
ones mentioned above exist in the industry and related
research works are available in the literature on informa-
tion sharing in the supply chain (Legner and Schemm 2008;
Chen et al. 2007), their focuses are mostly on the shar-
ing of demand information between various echelons in an
123
1084 J Intell Manuf (2012) 23:1083–1101
aggregated manner (especially between retailers and their
upstream suppliers). Furthermore, most of these studies lack
the ability to quantify the magnitudes of benefits although the
said benefits have been the center point for their discussions.
In order to understand the actual benefits of the information
sharing within a supply chain, the effect of this sharing should
be analyzed in conjunction with a particular control system

(e.g. hierarchical structure, heterarchical structure, or hybrid
structure) governing the supply chain. In this paper, we pres-
ent a coherent and comprehensive framework for analyzing
the impact of information sharing in and across various deci-
sion levels (e.g. enterprise level, shop level, cell level, and
equipment level) of the hierarchical, supply chain control sys-
tems. In particular, we consider a two-echelon supply chain,
where the hierarchical process plan selection from a set of
alternatives and real-time resource allocation are considered.
In this work, the primitive process plan for each level of deci-
sion is assumed to be available when a part production order
is submitted. These primitive process plans contain a com-
plete set of possible production alternatives including the
ones affected by the shop, cell, machine, and tool selections.
When the manufacturing system starts to operate, succinct
decisions have to be made on these primitive process plans
to determine which processes will be performed in which
sequence. First, the final serialized process plan is derived
from the primitive process plan considering the system sta-
tus. The making of these decisions is called the “linearization
of process plans”. Then, resource-allocated process plans
are generated by assigning the most appropriate resources
(i.e., shops, cells, machines and tools) to these linearized
process plans and sent out to the manufacturing system to be
executed.
Our particular objective in this study is three-fold. The
first objective is to refine a four-level hierarchical decision-
making framework to be used for the linearization of process
plans and real-time resource allocation in a supply chain sys-
tem, where the four levels are enterprise level, shop level,
cell level, and equipment level. The sources of information
affecting each decision level in the hierarchy are then iden-
tified in each of these levels. This four-level hierarchical
framework is constructed as a mechanism to enable opti-
mal or near optimal decision-making within the considered

supply chain. This optimal decision-making is enabled via
the multi-objective genetic algorithm, which is embedded
into our four-level hierarchical decision-making framework
mentioned above.
The second objective is to develop a novel approach to
investigate various roles of information sharing in manu-
facturing supply chains in a quantitative manner using a
test-bed developed in this work in conjunction with intelli-
gent decision-making algorithms. This investigation aims to
determine the magnitude of the impact of sharing of informa-
tion on linearization of process plans and real-time resource
allocation, where the overall goal is to increase efficiency
by decreasing the total cost of production and lead-time to
enhance its re-configurability. This is a unique study, which
allows us to simultaneously manage near optimal decision-
making via the considered four-level hierarchical decision
framework, operations based on the generated (linearized)
process plans and resource allocations, and automatic feed-
back that is used to check systems performance based on
these decisions. Given updated system status and require-
ments, the linearization and resource allocation of the process
plans will be regenerated if the values are beyond the allowed
threshold range, while specifically balancing the benefits of
such updates against the cost of schedule disruption. While
illustrated from the perspective of a single facility, the pro-
posed approach will permit modeling of large supply chain
instances with external suppliers and multistage fabrication-
assembly systems.
To this end, the critical aspects of the information to be
shared when developing models to help decision-making in
supply chain systems are analyzed in detail. These critical
aspects are classified in three main categories, including (1)
the axis of shared information, (2) the degree of shared infor-

mation, and (3) the content of shared information. The axis of
shared information can be either vertical, where the informa-
tion is shared among various hierarchical levels (i.e., cells)
within a single echelon (i.e., manufacturing facility) or hori-
zontal, where the information is shared across multiple ech-
elons of the supply chain (see Fig. 1). The content of shared
information refers to the actual information that is shared
between various members of supply chain and customers,
and can be defined both for vertical or horizontal axises. The
degree of information shared can be divided into three aggre-
gated levels for both axises: no information sharing, partial
information sharing, and full information sharing. In the case
of information sharing via the horizontal axis, no informa-
tion sharing defines the situation where the supplier only has
information on the orders received from the manufacturer and
must utilize historical data to augment the order information
when preparing demand forecasts. With partial information
Fig. 1 Information sharing among supply chain members
123
J Intell Manuf (2012) 23:1083–1101 1085
sharing, the demand distribution faced by the retailer and
the retailer’s inventory policy are known. Finally, with full
information sharing, the supplier also receives instantaneous
information on the retailer’s demand. In the case of infor-
mation sharing via the vertical axis, no information sharing
defines the situation where only shop-level aggregated infor-
mation is known and the operations planning framework has
no detailed information regarding various cells of a shop
floor or specific machines within these cells. In this case,
strategies can only be determined at the strategic level in a

highly aggregated manner. With partial information sharing,
the cell level information is known and shared within the
same shop floor. Hence, the planning can be performed at
the tactical level. Lastly, with full information sharing, all the
information generated from the smallest element of the shop
floor, which is considered as machine in this study, is known
and shared among other decision levels. Therefore, accurate
predictions as well as the decision tasks can be generated
for the operational level within this echelon of the supply
chain.
The accuracy of the information that is shared in differ-
ent dimensions mentioned above (i.e., capacity of machines
employed in the manufacturing process, shop floor layouts,
and maximum safety stock of the average on hand inven-
tory) is essential in generating feasible decisions in supply
chain systems. Distortion of this information could lead to
congestion and problems such as excessive work-in-process
inventory, back order costs, and increase in cycle time. The
consequences arising from faults at the shop floor level may
have detrimental effects on the performance of entire sup-
ply chain organization. Therefore, the impact of distortion
of information during the hierarchical decision-making has
also been studied, and solutions to mitigate the losses arising
from lack of or inaccuracy in information sharing have been
discussed in this research.
The third objective is to develop an accurate representa-
tion of a manufacturing supply chain involving a production
facility and a retailer facility as a simulated test-bed, where
realistic manufacturing scenarios can be mimicked to investi-
gate alternative “what-if” scenarios. These what-if scenarios
entail different shop, cell, and equipment configurations. This
test-bed is built in a great detail to allow for tracking even
very small structural changes (e.g. tool conditions, remain-
ing processing time at a machine) within the system and how

sharing of different information would affect the system per-
formance in a greater scale.
The remainder of this paper is organized as follows.
Section “Background and Literature Survey” summarizes a
background and literature survey related to this study. Section
“Proposed Approach for the Impact Analysis of Informa-
tion Sharing” gives an overview of the proposed approach
for the impact analysis of information sharing. Section
“Considered Manufacturing Supply Chain” presents the
details of the manufacturing supply chain system considered
in this work. In Section “Mathematical Formulation of the
Problem”, we first describe a generic mathematical formula-
tion of the linearization of process plans and resource allo-
cation problem for each level of the production system and
then explore ways of employing various parameters incor-
porating shop level, cell level and equipment level informa-
tion into this mathematical formulation. Section “Details of
the Multi-Objective Genetic Algorithm” explains the pro-
posed four level hierarchical decision-making framework
and its near optimal decision capability using multi-objective
genetic algorithm. Section “Experiment and Results” sum-
marizes the designed experiments together with the results
obtained from our case study. Lastly, in Section “Conclusion
and Future Work”, the conclusions are drawn and the future
aspects of this study are discussed.
Background and literature survey
Information sharing in supply chain systems
When information is shared in inter-organizational networks,
it can result in a more efficient flow of goods and services,
reduced inventory level, and lower costs, which benefits the
overall network (Venkateswaran and Son 2004). Information

sharing has also been found to reduce the bullwhip effect
(order variability) and to add value to the chain (Chen et
al., 2000; Lee et al. 1997). For example, Wal-Mart shares
point-of-sale information with their suppliers and transmits
orders electronically to the relevant supplier when inventory
for an item falls to a predetermined minimum level of stock
(Lancioni et al., 2000). This inter-organizational information
sharing reduces carrying costs of inventory, facilitates quick
response for inventory replenishment, and allows suppliers
to better plan their production schedules, and reduces lead
times (Stevenson, 1994; Baganha and Cohen 1998; Sherali
et al. 2008).
While benefits of inter-organizational information shar-
ing has been widely studied (Chen et al. 1998; Lee et al.
2000), it is difficult for firms to reap these benefits unless
they have a better grasp of the antecedents of inter-organi-
zational information sharing that influence its effectiveness.
This is an area in which firms need guidance so that they can
effectively channel resources to their knowledge and infor-
mation sharing activities. The ability of managers to coordi-
nate the complex network of business relationships that exist
between parties involved in the supply network is critical to
the success of a firm (Lambert and Cooper, 2000).
The network patterns, which can have one or more stages,
create varying levels of complexity, and hence differing envi-
ronments for information sharing. Further, location of the
firm(s) within a network creates different information needs
123
1086 J Intell Manuf (2012) 23:1083–1101

and consequences due to the distortion in the flow of informa-
tion. Managing the flow of information effectively requires
close attention to the coordination mechanism established
among the member firms in a network. The degree of coor-
dination is likely to affect the nature and amount of informa-
tion that is shared across the network. Similarly, the degree
to which the partner firms perceive a match in their goals
may impact the nature and amount of information they are
willing to share with each other (Samaddar et al., 2006).
Lee et al. (1997) explain the importance of mitigation of
detrimental impact of distortion of information in the form
of inbound orders, from downstream members to upstream
members, in a supply chain. They consider a series of compa-
nies in a supply chain, and each of the companies orders from
its immediate upstream member. They claim that the distor-
tion is amplified in the upstream direction of the supply chain
as a demonstrative of the “bull whip” effect. They analyzed
demand signal processing, rationing game, order batching,
and price variations as four sources of the “bull whip” effect.
In particular, there is an amplification and larger variance in
the distortion of information from downstream to upstream
members. Sterman (1989) attributes the increase in amplifi-
cation and variance of information distortion to several “mis-
perceptions of feedback”.
As mentioned earlier, a great percentage of the literature
examines the impact of information sharing in the supply
chain in a highly aggregated manner without considering
the possibility that the organizations may benefit much from
the information sharing in a detailed level. However, shar-
ing of information horizontally among the partners of sup-
ply chain (i.e., demand, work-in-process levels, replenish-
ment time and amount) and then sharing the elements of
this information vertically within a production facility (using
hierarchical decision-making process proposed in this work)

may help organizations work in their highest efficiency. Sim-
ilarly, sharing of data originated from the production floor
(i.e., cell, and equipment related data) with various members
of the facility may enhance the productivity and efficiency
level leading to minimized lead times, work-in-process, cost
and therefore, maximized profit. In this work, we intend to
develop a coherent and comprehensive framework for ana-
lyzing the impact of information sharing among various deci-
sion levels of manufacturing supply chains.
Hierarchical decision-making
In a hierarchical decision-making system, the commands typ-
ically flow from top to down, and feedback information flow
from bottom to up, and the relationships among the modules
are limited to master-slave relationships between parent and
child nodes in a tree shaped hierarchy (Bongaerts et al. 2000).
One common decision area considered in the literature within
this natural hierarchy is process planning. Process planning
is the task of planning the various processes or operations to
be made on raw materials to transform it into the desired part.
The output of this task is a plan that contains the routes, pro-
cesses, process parameters, machines, tools and set-ups to be
used for manufacturing parts (Ma et al. 2000; Nau and Chang
1983). While computer aided hierarchical process planning
has advanced significantly in its duty of creating the prim-
itive process plans with alternatives, human intervention is
still needed as decision mechanisms to finalize process plans
by serializing the process plans after considering alternate
sequences of the operation to be performed, and then allo-
cate the resources based on this final process plan. These
decisions are quite important as they are essential for pro-
duction and determine the efficiency of production (Joshi
et al. 1986). Other approaches (Dilts et al. 1991) have used
the hierarchical decision-making architectures to operational

decisions via off-line schedulers at different levels of the hier-
archy. In these approaches, when a disturbance occurs, it is
fed back to the appropriate level in the hierarchy (e.g., the cell
scheduler), and only after the schedule has been adapted, the
new schedule triggers a new flow of commands that forms
the reaction to the disturbance. The resulting response time
is low leading to a low robustness (Bongaerts et al. 2000).
In this study, we propose a multi-objective genetic algo-
rithmic approach to be used as a decision aid for hierarchical
linearization of process plans and real-time (on-line) resource
allocation inside a production facility of a two-echelon sup-
ply chain and help analyze the effect of information shar-
ing in manufacturing systems. Real-time resource allocation
decisions at each level determine which resources will be
utilized for the production. For instance, our considered pro-
duction facility comprises two shop floors, each divided into
cell areas that are comprised of machines and tools. Here,
our resource allocation decisions take into consideration the
resource availabilities (e.g., shops, cells, machines and tools)
as well as the part routing through these resources. Follow-
ing the assignment of jobs to a specific cell at an aggregated
level, the machines are allocated to specific operations within
these jobs based on the processing cost and times of these
operations in a more detailed level. This hierarchical nature of
allocating resources at a lower level after allocating resources
at a higher level applies to all levels of decision-making.
Meta-heuristics
Process plans that are selected without considering the
resource availability (which is imminent in the operation plan
of the whole production unit) tend to violate feasibility con-
straints and lead to losses. Efforts to find the optimal or near
optimal solution for this problem have been spent mainly on
two different approaches. In the first approach, the linear-

ization of process plans (selection) and resource allocation
problems are combined and formulated as a single problem
123
J Intell Manuf (2012) 23:1083–1101 1087
(Moon et al. 2002; Yan et al. 2003). The second approach is
comprised of two steps, where the first step (Aldakhilallah
and Ramesh 1999; Sormaz and Khoshnevis 2003) involves
the generation of numerous process plans that achieve the
predetermined level of performance criteria in the process-
planning problem. Then, in the second step, optimization of
the scheduling problem is performed with respect to each
of the alternate process plans generated above, where each
imposes different feasibility constraints.
The search space for the combined linearization of pro-
cess plans (selection) and resource allocation problems is
exhaustive, which makes it difficult for deterministic opti-
mization methods like integer programming or mixed inte-
ger programming to arrive at an optimal solution. The said
search space grows even bigger and makes the problem even
harsher to resolve when the problem is formulated in detail
for large-scale complex systems such as the considered sup-
ply chain system in this research. In order to address these
challenges, another type of approaches have emerged through
the use of meta-heuristics. Meta-heuristics including genetic
algorithms (GA) (Moon and Seo 2005), simulation anneal-
ing (SA) (Cai et al. 2009), particle swarm optimizations
(PSO) (Guo et al. 2009), and Tabu search (Brandimarte and
Calderni, 1995) have been employed to obtain solutions to
problems which belong to the class of integrated process
planning and scheduling (IPPS). GA, evolutionary algorithm

(Li et al. 2009), and symbiotic algorithm (Kim et al. 2003)
have all been quite successful in obtaining near optimal solu-
tions to variants of IPPS problems. The concept of reaching
the optimal solution by a way of evolutionary synthesis of
the pool of candidate solutions is common to GA, symbiotic,
and evolutionary algorithms. In this study, we have designed
a multi-objective genetic algorithm in order to take advan-
tage of the method of reaching a near optimal solution via
evolution in reasonable short time.
Proposed approach for the impact analysis of
information sharing
Figure 2 depicts a hierarchical decision-making framework
considered in this research, in which we intend to quantita-
tively analyze the impact of information sharing in manufac-
turing supply chains. The framework consists of enterprise,
shop, cell and equipment levels. There exists a controller
(including monitoring, decision-making and execution mod-
ules) at each level, where its decision-maker could be based
on a separate multi-objective genetic algorithm depending
of the complexity of its decisions. Each controller is associ-
ated with one resource at the same level, and interacts with
the controllers at their immediate lower levels. These inter-
actions include (1) the information shared about the status
of resources at a lower level by controllers at that level,
(2) the decisions to lower level controllers from higher level
controllers, and (3) the feedback information on the perfor-
mance of resources at a lower level. After process plan lin-
earization and real-time resource allocations are received at
the system monitor, they are processed by the decision-mak-
ers, and the resulting decisions to the lower level controllers
are communicated via the executor. We assume that there
is no direct interaction between any two controllers, other
than those at successive levels, for the ease of analysis of

International Journal of Production Research,Vol. 45, No. 11.docx

International Journal of Production Research,Vol. 45, No. 11.docx

Recommended

Recommended

More Related Content

Similar to International Journal of Production Research,Vol. 45, No. 11.docx

Similar to International Journal of Production Research,Vol. 45, No. 11.docx (20)

More from normanibarber20063

More from normanibarber20063 (20)

Recently uploaded

Recently uploaded (20)

International Journal of Production Research,Vol. 45, No. 11.docx