This document discusses production planning and control (PPC) systems. It examines the current state of PPC and how PPC systems are being asked to respond more dynamically to changes in technology, markets, and customer expectations. The document provides an overview of what PPC is, including defining key aspects like the master production schedule and hierarchical planning processes. It also discusses some of the challenges in practical PPC implementation and controlling complex, globally dispersed supply chains.
Parag Bhagat - The Management Accountant - Oct 2015
Reflections inproductionandplanning
1. REFLECTIONS ON PRODUCTION
PLANNING AND CONTROL (PPC)
Invited Paper
Maurice Bonney
School of Mechanical, Materials, Manufacturing
Engineering and Management
University of Nottingham – UK
v.7, n.3, p.181-207, dez. 2000
Abstract
The paper examines the current state of Production Planning and Control (PPC), identifies some
technical and systems changes that have occurred over recent years and links these with the require-
ments being placed on companies by the market. PPC is being asked to respond effectively to these
internal and external changes by being more dynamic and providing better control of resources and
delivery performance. Some of the requirements to be satisfied by the new PPC systems are identified.
To meet these requirements it is suggested that better understanding is required of how different
factors affect PPC systems performance and that administrative systems need improving. The
quantitative, administrative and behavioural aspects of PPC are discussed. A framework for develop-
ing an agenda for action and research is provided.
Key words: production planning, production control, overview.
1. Introduction Production Planning and Control (PPC) is being
asked to respond effectively to these internal and
M any technical and systems changes have
occurred in manufacturing industry over
recent years. The requirements being placed on
external changes by providing a faster response
and better control of resources and delivery
performance.
companies by the market are also changing. The paper examines the current state of PPC.
2. 182 Bonney – Reflections on Production Planning and Control (PPC)
The discussion then reviews recent develop- to be critical of the content and to ask whether
ments in the market, manufacturing and the interpretations are consistent with readers’
manufacturing systems and relates these changes experiences. Feedback on whether the generali-
to our understanding of production planning and sations stand up to international scrutiny and
control. The description brings together many apply to different systems will be useful.
threads associated with PPC and identifies some
areas where there are deficiencies in our 2. What is PPC?
understanding. Some thoughts are then presented
on how PPC systems need to respond to the
changing technology, changing market needs
and individual customer’s expectations.
C ompanies wish to satisfy market demands
expressed in terms of real or forecast
demand. To do this companies in general
All papers are personal in the sense that they produce a Master Production Schedule (MPS)
reflect the interests and priorities of the authors. that states the number of each product to be
However personal opinions have deliberately made over some planning horizon and a Sales
been given greater weight in this paper than is Programme that states the number of each
conventional. Indeed, several sections of the product to be sold. The PPC function and its
paper contain ideas that have been partially associated systems aim to plan and control
tested but not fully evaluated. In that category production so that a company meets the
are the personal classification and organisation production requirements as effectively as
of PPC described in section 4, the concurrent possible. PPC systems are hierarchical. A
engineering views and figures in section 7, the hierarchical planning process is used for PPC to
framework for production management in help a manager understand and control the
section 8.1 and the tabular representation of operations for which he is responsible. A high-
actions in section 9. How the tabular representa- level plan sets the context within which the next
tion leads to a possible research agenda also lower level plans operate. The different levels in
needs further evaluation. the hierarchy operate on different time scales.
The paper uses some general principles Typically aggregate planning is associated
which, although not proven, have been rein- with long planning horizons and detailed
forced over the years by personal experience and planning is associated with short planning
research investigations. Hopefully these principles horizons. A common operational starting point
will also be useful to you. They include: for planning production is the Master Production
• theory, simulation and practice should Schedule from which the requirements of
complement each other; materials, parts, machines and labour are
• systems integration is important; derived. Modifications may be made to the plans
• system commonality and structures are a if the derived requirements are thought to be
useful basis for transferring experience; inappropriate. The consequences of the chosen
• people are an integral and essential part of plans, usually expressed as a list of requirements
systems; of made-in and bought-out parts, become the
• automatic provision of data can greatly assist basis of the work schedules placed on individual
system maintenance; men and machines and orders placed on
• information is important; suppliers.
• provision of PPC systems or information The prime objective of production planning
systems should be based on a proper eco- and control is to ensure that parts and products
nomic assessment. are produced so as to achieve the Master
The conjectural nature of some aspects of the Production Schedule (MPS) in a way that is
paper has been highlighted to encourage readers consistent with meeting the company’s other
3. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 183
performance measures. The MPS states the putting effort into reducing or even eliminating
number of each product to be produced period the elements of complexity and uncertainty
by period over some time into the future known inherent in manufacturing operations…’ Factors
as the planning horizon. The MPS, in conjunc- mentioned include: design for ease of manufac-
tion with the sales programme, is the company’s ture, improving factory layout, introducing
planned response to the demands of the market. improved production methods, improving
In particular, the MPS, company production and quality, etc. Some of these are discussed later.
inventory policies and knowledge of inventory
levels, are used to determine the number of items 3. The Current State of Practical PPC
to produce, the planned inventories of raw
material, work-in-progress, finished parts and
finished products and the manufacturing
resource requirements such as machines and
P PC systems consist of inputs, transforma-
tions, outputs and appropriate control
systems. Traditionally it has been difficult to
labour. Plans should be realistic and avoid plan and control production because of limita-
asking for the impossible. tions in the planning methods and because
The British Standards document BS 5192 unplanned changes occur in demand, supply and
Part 1:1993 states that the Production Control resources. In recent years further complexity has
function comprises three inter-related stages: arisen because the emphasis of PPC has shifted
programming, ordering and dispatching. It also from controlling individual plants to co-
states that production control occupies a central ordinating the complete supply and delivery
position in the exchange of information between chains. The production units may therefore be
the functional departments within a manufactur- geographically dispersed and the systems in the
ing organisation. In particular it then goes on to different units, which are not necessarily part of
mention the following seven groups of functions: the same company, may not integrate easily with
sales/marketing, design/development, purchas- the others. The consequence is that many
ing, finance/accounting, manufacturing/quality companies find it difficult to provide the service
assurance/production engineering, distribution, required by the market. They also find it difficult
personnel. A common representation of PPC to produce items according to their plans despite
includes the following characteristics: major developments in computer aided
• A hierarchy of planning that involves a production planning and control and better
progressive detailing of high level plans to systems understanding. Difficulties of planning
produce operational plans and associated and control arise when there is a mismatch or
instructions; incompatibility between various parts of the total
• Communication that allows the plans to reach system i.e. any of the following:
the appropriate people at an appropriate time • The input (the demand);
and • The desired or chosen output (the Master
• Feedback that provides suitably summarised Production Schedule and its implications in
information about performance to the control- terms of production plans and plans for the
lers of the plans. supply of parts and material);
• The chosen production plans;
BS 5192 also makes the points that ‘there is a • The supply system that should ensure that
tendency when designing production control suppliers deliver the correct quantities of
systems to make them over-complicated.’ and assemblies, parts and materials of appropriate
that ‘In many situations there is an alternative to quality on time;
developing ever more complex systems to deal • The conversion process i.e. the production
with increasingly complex operations. It involves system and how well it performs;
4. 184 Bonney – Reflections on Production Planning and Control (PPC)
• The control of the conversion process (how companies have local objectives such as keeping
deviations from the production plans, system groups of individual men and machines busy,
performance, forecasting system and supply formal descriptions of PPC systems usually sub-
system are dealt with). divide them into make-to-order and make-for-
stock systems. Make-to-order systems, of which
There is frequently a gap between the theory jobbing production is the purest manifestation,
and practice of PPC. Academics try to improve respond directly to customer’s demands. On the
understanding of PPC systems by analysing the other hand, make-for-stock systems, including
mutual impact of influential PPC factors base stock, re-order cycle (ROC) and sometimes
whereas practitioners try to obtain usable results re-order point (ROP) systems, attempt to bring
from a mix of non-ideal software and ad hoc stock up to pre-defined levels. Many systems,
manual systems. Practitioners often believe that of which MRP is one, are a combination of
academics are not investigating the ‘right’ make-to-order and make-for-stock. In MRP,
problems and are touched by the arrogance of customers’ orders are accepted within the
ignorance whereas academics are frequently constraints of the Master Production Schedule
unhappy about the apparent lack of understand- (MPS). Conversely the MPS is chosen so that
ing that production control managers have of orders that have already been accepted will be
straightforward concepts. As an illustration, produced. Of recent years, Just-in-time (JIT),
practitioners need to agree lead-times at the time which according to APICS (1998), is usually
of accepting orders so as to be able to quote considered to be a philosophy based on the
dates for delivery and to co-ordinate the ‘planned elimination of all waste and continuous
availability of parts for assembly. Selecting a improvement of productivity’, has become
delivery date is usually achieved by assuming fashionable and sufficiently important to become
that the lead-time is fixed and deterministic. This an objective of many companies. JIT is usually a
may be a reasonable approximation provided make-for-stock base stock system suitable for
that the load on resources stays fairly constant or repetitive manufacturing, but it has two features:
fairly light. However, queuing theory clearly demand pull and the minimisation of stock that
demonstrates that highly loaded resources, in the together allow the system to respond effectively
presence of demand and service both subject to to customer orders. JIT frequently operates
variability, create long and variable delays. within the context of requirements derived by an
Actual lead-times are thus a consequence of the MRP system. JIT has had some important
load on the system together with random effects successes but it has not been universally
arising from problems of quality, reliability of successful. Indeed the effectiveness of JIT
the machines and processes, deliveries, people systems has frequently been somewhat different
and the demand. Rather obviously the quoted from that expected even in companies that
lead times affect the load, the level of stock and have claimed success in its implementation e.g.
the probability of delivering on time. However, see WHYBARK (1996) and VASTAG &
companies frequently adopt simultaneously the WHYBARK (1993).
contradictory strategies of varying load while MRP is still the most used production plan-
fixing the quoted lead-time. This may be one ning and control system. Over the last twenty
reason why many plans are unrealistic. years the procedures of MRP have been
There are many systems of production extended and developed to create Capacity
control. In general, PPC systems have originated Requirements Planning (CRP), Distribution
within industry although the major system Requirements Planning (DRP), Manufacturing
software producers now have a great influence Resource Planning (MRP II), Enterprise
on their detailed content. Although some Resource Planning (ERP), etc. However, the
5. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 185
basic system embedded within these extensions, Instructions are issued and transactions related to
is still MRP, an approach that has been heavily materials, men and machines are recorded.
criticised partly because its centralised nature Unless resources are lightly loaded, it is
requires a large amount of data handling and the common to find that order acceptance is not
need to maintain highly accurate stock records. compatible with the resources available. This
Nevertheless, it is essential for companies to co- introduces uncertainty into the delivery times on
ordinate deliveries and a wide range of other top of the uncertainty that arises from changing
activities and so MRP has continued to be used customers’ requirements, machine breakdowns,
even though the data handling may preclude absenteeism, quality problems, etc. The
responsiveness. The choice of parameters consequence is likely to be a set of dissatisfied
available within an MRP system and the customers. To try to overcome these problems,
differences between companies and their procedures are introduced to perform actions
markets mean that it is unlikely that any two such as master scheduling, lot sizing, setting
MRP systems are identical even if the companies safety stocks, rescheduling, setting time fences
use the same software. JIT and kanban systems and extending planning horizons. Computer
also have great variety in the way that they have systems are introduced to handle the required
been implemented. In practice the same label, data more effectively. While these methods help,
whether MRP or JIT, may be used to describe they do not completely meet the needs of the
systems with different structures, different companies and they frequently do not meet the
parameter values and very different perform- needs of the users. Together the procedures used
ance. More confusing, but less frequently for planning and control make PPC systems
recognised, is that the labels, push and pull, that surprisingly complex.
are commonly used to describe MRP and JIT Since the earliest days of computerised
systems are also defined in many different ways. production control, an objective of companies
Thus not only can different systems be described has been to develop Integrated Data Processing
in the same way but the same systems may also (IDP) systems. From the late 1950’s to mid
be described in contradictory ways. This has 1970’s my personal experience in companies
been discussed in BONNEY et al. (1999a) and included linking financial and administrative
that same paper reports a simulation study that systems with PPC, and linking CAD and PPC. In
compares the performance of very simplified each case the systems formed the basis of a
push and pull systems. The simulation indicates company wide Management Information System
that under some specific conditions, push (MIS). In those early days bespoke PPC systems
systems, appropriately defined, have similar and modular packages both existed. TATE
potential for stock reduction as pull systems. The (1976) showed that the early modular packages
lack of precision with definitions has led to were not as successful as bespoke systems.
confusion within the literature and among However, the use of packages has steadily
production control practitioners. This suggests increased so that the practice of MRP is now
that to improve understanding it is essential to dominated by modular software. In recent years
represent systems with clarity and desirable to several big software houses have taken a major
represent them consistently. share of the market although there are still many
PPC systems in companies frequently pro- smaller software/consultant companies imple-
gress through a series of stages. In the early menting successful systems which apparently
stages most company PPC systems are adminis- perform as well as the systems from the major
trative systems built up ad hoc to deal with the players. With nearly every modular system,
housekeeping aspects of planning, ordering, users are able to choose the modules and the
receiving, issuing, making and storing items. sequence in which the modules are implemented
6. 186 Bonney – Reflections on Production Planning and Control (PPC)
into the company. The values of the parameters, manufacture, by quantifying many of the
used by the modules such as planning horizons, ‘intangibles’. It is particularly important that
batch sizes, planned safety stock levels, clear objectives are set when planning the
smoothing constants, etc. are also chosen by the system introduction so as to ensure that the
users. Of these it is likely that the time parame- expected gains are actually attained.
ters (planning horizons, time buckets, time Surprisingly, many of the mistakes made in
fences etc.) are the most critical. In many cases the early day of implementing computer systems
the values chosen are based on company still persist and the chosen implementation time
convention rather than on an understanding of scales are frequently optimistic. The most
how these parameters affect overall perform- important requirement is that the top manage-
ance. Probably because the potential gains that ment is committed to the project and that
can be obtained from the use of ‘good’ PPC appropriate preparatory work is done before
systems are not generally appreciated, the going live. The preparatory work must ensure
integration of the company’s administrative that staff are trained, appropriate data are
systems is more likely to have a higher priority available, the software is usable and that
than the improvement of PPC performance. The operational trials are undertaken. Whenever
criteria by which a PPC system should be judged practicable the trials should include running the
e.g. delivery performance, quality expectations, proposed and current systems in parallel.
cost, service, etc., are not fully agreed, perhaps
because they are dependent on market expecta- 4. Classification and Organisation of PPC
tions or competitors’ actions.
Performance of PPC systems is improving in
the sense that lead times are reducing and
delivery performance is improving. However
P PC systems can be classified, structured and
organised in many ways. For the purposes of
this paper, even though the categories are not
this has been achieved mainly by administrative independent, systems are somewhat unconven-
means rather than by a clear understanding of tionally listed under the following headings:
the relationships between the variables. For structure, mathematical modelling, data control,
example, as a result of using better data or human centred systems and managerial style.
performing simple arithmetic calculations, it Invariably advantage may be gained by
may be obvious that there is a mismatch between simplifying the organisation and improving
different parts of the system e.g. the resources procedures. Items of this administrative nature
required and the resources available. If so, then are listed as a sixth category. These categories
action can be taken to increase or reduce the are now discussed.
resources. In some cases, system performance System structure can be the basis of classify-
may be improved easily e.g. by producing better ing PPC systems. Any structural representation
demand forecasts, improving the master needs to recognise that production planning and
scheduling procedures or improving inventory control is hierarchical and that different levels in
recording. In other cases, the company may need the hierarchy operate over different planning
to wait for an appropriate time to make the horizons. Structures need to be consistent with
necessary investment e.g. in new machinery. developments in information technology and
There are frequently difficulties in meeting should also be able to represent the dynamics of
market demands within the constraints imposed PPC systems. The GRAI model (DOUMEIGNTS
by conventional investment appraisal criteria. et al., 1992) was developed primarily for systems
However PRIMROSE (1991) suggests ways of analysis and design whereas the aim of the
appraising advanced manufacturing technology framework for production control (BONNEY &
capital projects, including MRP, JIT and cell HEAD, 1993 and BONNEY et al., 1999b) has
7. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 187
been to increase understanding and assist the company or they may have been recently
process of designing production planning and introduced or proposed functions resulting from
control systems. Some of the advantages of the say a Business Process Re-engineering (BPR)
framework and Petri-net modelling are discussed exercise. Data control systems are straightfor-
in Section 8.1. Petri-net modelling and control ward to represent by IDEF diagrams. Emphasis
modelling, both useful representations for on data control is often implicit within computer
modelling the structure and the dynamic based control systems but there is a danger that
performance of PPC, are discussed respectively an analyst will ignore the needs of some users
in Sections 8.1 and 8.2. and that some discretionary decisions will not be
Representing systems by mathematical models acknowledged. Such problems often come to
allows the performance of a PPC system to be light only when the new system becomes
derived as a consequence of changing parameter ‘operational’.
values in the mathematical model. The approach Human centred or behavioural views of PPC
is an effective way to improve understanding. In systems are now recognised as important.
principle, if a mathematical representation was Human centred systems emphasise the needs of
realistic enough it could be used to control the the system users. Most studies of PPC take either
operation of the system. However, in practice a human centred view or a deterministic view,
there is a great deal of discretionary decision- but means are needed to combine the determinis-
making in the operation of systems involving tic and the human centred views. There are so
people. In particular, differences in people’s many uncertainties in every production planning
cognitive processes are hard to represent and control system that a human centred view is
precisely. It is therefore unlikely that mathemati- essential to allow the people responsible for
cal models will ever control complete plants controlling the flow of work to be able to
except possibly in the process industries. In contribute to the decision making. For example,
more traditional batch or jobbing manufacture, as described by MACCARTHY et al. (1999),
mathematics may be used to represent near even in areas such as scheduling, a problem that
deterministic situations e.g. to derive a master is often expressed as a classical mathematical
production schedule or to optimise some problem, human discretionary decision making
operational sub-parts of the problem. However, is essential. People understand the current local
even problems with a classical structure e.g. difficulties whatever their cause. However, in
sequencing operations on the shop floor, require order to contribute effectively, they will require
close co-operation between the user and the appropriate information presented in a form that
algorithms to ensure that the discretionary needs they can understand. Ideally they need an
of the users are not ignored. As system perform- appreciation of causality and knowledge and
ance becomes more deterministic, the easier it experience of constructive methods of conflict
becomes to represent mathematically and for its resolution.
operation to become computer controlled. Systems based on managerial style are sys-
However, in most situations that prospect still tems where the personality of the manager
looks a long way off because, in addition to the together with his cognitive style is a major
conditions not being sufficiently deterministic, component in getting the system to produce the
many of the problems are still intractable. results required. Direct managerial control is a
Data control systems consider a system in natural and effective way to react to crisis
terms of data and its flow between the different conditions and operate in highly dynamic
activities performed. The activities can be markets. Unfortunately this style of system
manual or computer based functions. They may operation often carries over to different
be long standing activities performed by the conditions. Managers may continue to react to
8. 188 Bonney – Reflections on Production Planning and Control (PPC)
current crises and events even when these are the regular schedule means that everyone knows
consequence of lack of planning rather than what is to be done without further instruc-
inherent conditions of the market. Resources tions;
may then be used inefficiently and ineffectively • having spare capacity as a contingency to
as the systems continue unnecessarily to be reduce the need for formal scheduling;
dominated by fire fighting activities. Obviously, • holding stock so as to provide a buffer
the effectiveness of reacting rather than planning against changes in demand.
is dependent on the stability of the market
conditions but, whatever the state, attempts to 5. Requirements Arising from Changes in the
formalise the processes are a natural develop- Market
ment. Anecdotally many personality driven
systems that champion leadership ignore the
needs of colleagues.
Organisational and procedural actions that
W henever and wherever companies innovate
with products, processes, systems and
marketing, that, in turn, imposes requirements on
simplify problem structure and improve other companies to change e.g. to redesign
performance can be associated with any or all of products and shorten delivery lead-times. Low
the above. They are the administrative changes wage economies force industrialised nations to
discussed later. Typical changes that may withdraw from the market, to sub-contract or to
simplify manufacturing control include: maintain their competitiveness by investing in
• formalising procedures; productivity and in product and process
• having someone in charge; innovation. The market wants low cost, high
• ensuring that data is accurate and up to date; quality products available within a short delivery
• changing factory layout to improve material time. Companies wish to have high productivity,
flow; short manufacturing lead-times and to minimise
• introducing product organisation and using stocks. They also need flexibility so as to be able
cell manufacture and GT (Group Technol- to respond easily to changing requirements.
ogy) for part manufacture and assembly, Environmental, health and safety and product
avoids many routing and control problems by liability legislation impose other requirements on
requiring all items to follow the same se- the organisation e.g. the need to be able to trace
quence; parts within a product. Environmental factors
• using methods of continuous improvement; add the need to reduce the amount of packaging,
• reducing unplanned activities and introducing to design for disassembly, to maximise the
lean manufacture methods e.g. by improving possibilities for recycling materials, to use more
quality progressively to producing zero environmentally friendly materials and processes
defects, improving reliability of machines, and so reduce pollution and energy expenditure,
reducing the involvement of people by using etc. The need to be able to maintain products in
robots to automate production, etc. avoids locations determined by the customer adds
unnecessary disruption; another dimension to design. There is also
• introducing JIT pull systems that enable the increasing concern about social issues such as
company to respond directly to customers the effects of shift work, the distribution of tasks
demand; between workers, designing jobs and workplaces
• sub-contracting or moving to a ‘virtual to meet the needs of specific populations
enterprise’ to avoid many of the internal including different ethnic backgrounds, disabled
control problems; workers, different sexes and different age
• working to a fixed programme despite distributions. Many other examples of changes,
fluctuations in the demand and working to a legislative constraints and the effects of
9. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 189
globalisation could be added. In particular many more secure and reduce the time and cost
problems are compounded by changes in the of production;
relative power of companies arising from − Using concurrent engineering to perform
mergers and take-overs and the changing values design and process planning, manufactur-
of currencies. ing systems design and PPC system de-
sign in parallel. This is discussed in
6. Requirements Arising from Changes in greater detail in Section 7.
Manufacturing and Manufacturing Systems
• Changes to hardware, production methods
M anufacture is responding to the needs of
the market in a variety of ways. The
responses include changes to the design process,
and production support. These include:
− Using different processes, more reliable
and faster conventional machines;
changes to hardware, production methods and − Using automatic methods to make indi-
production support, changes that provide data vidual components and assemblies more
more quickly and of better quality, organisa- quickly, reliably and flexibly;
tional changes, changes to planning and control − Implementing Computer Aided Manufac-
methods, etc. All of these changes need to be ture (CAM) including CNC, DNC, ma-
assimilated and integrated so that the total chining centres and robotics;
system remains a coherent integrated whole. One − Using cell manufacture;
way to achieve this is to construct the system − Using conveyors and storage systems to
from a flexible set of customised modules provide better material flow and control;
attached to a software platform. In principle an − Using different external transport to speed
appropriate software platform should allow up receipts from suppliers and deliveries
systems to be changed quickly and without much to customers.
disruption. The listed categories are not mutually
exclusive in their effects e.g. different planning • Changes that provide better quality data more
and control methods can improve the quality of quickly. These include:
data. Conversely providing data more quickly or − Using computers to store and process data.
of better quality may allow production planning − Implementing CAD to design products
and control to be improved. Some examples and to provide and maintain the bills of
within each category of change are: materials and the associated product struc-
tures data that will also be used by the
• Changes that may improve the design process PPC system;
and shorten the design time. These include: − Implementing CAPP to determine the
− Using computer-aided design (CAD) and processes to be used and the time they will
computer-aided process planning (CAPP) take;
to improve the consistency and flexibility − Implementing computer aided work-study
of the design process, to reduce the design to determine the methods and times re-
lead-time and to make design modification quired to make a product using manual
procedures more straightforward; methods. This is the equivalent to CAPP
− Using rapid prototyping to reduce the time for manual work and so in principle it is
to produce prototypes; now possible to provide the planned op-
− Designing for ease of manufacture (DFM eration times for all work before the work
and DFA); starts;
− Using modular assemblies so as to reduce − Using bar code readers and sensors to
variety in the organisation, make costing collect data more quickly and reliably. Bar
10. 190 Bonney – Reflections on Production Planning and Control (PPC)
coding may be used as the basis for col- quences of changing system parameters are
lecting data that describes the current clearly understood. When change was relatively
status (quantity, quality, location, stage of slow, experience was a reasonable basis for
production or test etc.) of all aspects of the action. However in times of rapid change,
production system. The data can relate to conceptual understanding becomes progressively
material, parts, products, deliveries, re- more important and trial and error, which is
turns, etc. and also to operators, machines the basis of experience, is too slow and risky.
etc. A separate decision is needed to de- Better conceptual understanding and improved
termine whether collecting this data is operational performance of PPC is particularly
economically desirable; important because it drives much of the rest of
− Using e-commerce to speed up the order- the organisation.
ing and delivery process. Companies need to plan and co-ordinate all
functional stages related to introducing a product
• Changes to the organisation of manufacturing from the initial product design, through produc-
systems. The company changes could include: tion, distribution and sales and after sales service
− Changing objectives, goals and strategy to scrapping and possible recycling. They also
arising as a result of management changes need to choose and create the associated systems
and reviews; that enable the stages to be planned, co-ordinated
− Using more focussed, flatter management and controlled. There is a need to provide better
structures and reduced vertical integration; responsiveness to the external changes discussed
− Extending control to the supply chain; in Section 5 including competitors’ actions and
− Making organisational and procedural changing market demand. This means that
changes to simplify problems as discussed companies should be able to reduce or increase
in section 4. production levels without ceasing to be profitable,
to change models to meet new perceived needs,
• Changes to production and inventory to change systems without excessive consequen-
planning and control techniques and tools. tial costs, etc. It is desirable to reduce lead times
These include: while maintaining or even improving service
− Improving master scheduling and its use; levels. This requires a better understanding of
− Implementing computer aided production the consequences of choosing specific values of
management (CAPM) software such as the parameters such as safety lead times,
material requirements planning (MRP) frequency of scheduling, effect of delays and
and using better scheduling methods. This inaccuracies of recording, etc. It probably needs
can help to plan and control production; the dedicated development of lean manufactur-
− Using simulation to investigate many ing methods (total quality, dedicated workforce,
aspects of the production system before reliable and accurate machines, reduced set-ups
the work starts and to investigate the con- and small batches). There is a need to develop
sequences of a proposed production pro- better ways of evaluating investment in
gramme and its detailed schedules. appropriate capacity. PPC also needs to respond
to the changes described in section 6. The high
7. Consequential Production Planning and rate of change that has been described suggests
Control Needs that there is a need for routine procedures that
systematically adjust the system consisting of
P roduction planning and control technology
needs to change from a craft process into a
properly engineered process where the conse-
products, processes, manufacturing system
design, manufacturing and PPC so that they
remain in balance.
11. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 191
Market Requirements
Company
requirements Unsolved problems
(internal)
Product
Product specification
Design
A1
Unsolved problems
Process Product and process
Design specification
A2 Unsolved problems
Manufacturing
Systems Product, Process
Design & System Specification
A3
System
Design Designed
A4 Systems
Figure 1 – A concurrent enterprise representation.
The paper has illustrated that manufacturing system design’, ‘business system design’ and
and logistics systems are complex, integrated ‘external systems design’ as shown on Figure 1.
and dynamic. Changes to any of the functions Applying these concurrency ideas to the whole
e.g. product design, process design or manufac- organisation defines the concept of a ‘concurrent
turing systems design require adjustments enterprise’. Figure 1 is a structured representa-
elsewhere. A method of representing the tion that links product, process and manufactur-
dynamics of the inter-relationships between the ing system design with ‘system design’. Figure 2
different functions is important, as it will help to is a more detailed representation of the process
clarify the relationships. In BONNEY et al. of determining the ‘PPC system’ part of ‘system
(2000a), the product design, process design and design’. Figure 3 represents the planning
manufacturing system design stages of introduc- hierarchy of an MRP system, which is a possible
ing a product into production are represented in output of the design process illustrated in
structured analysis format. The stages are kept in Figures 1 and 2. These ideas were discussed in
balance by the use of a hierarchical, iterative BONNEY et al. (2000b), which contained much
design process. Detail may be progressively of the information of sections 5 and 6 and also
added to each stage of the hierarchy but if the listed some of the major changes facing
detailed consequences are unsatisfactory it is production organisations from the market and
possible to return and modify the previous stage changing technology. It then presented in an
before adding further detail. The need to almost random sequence a list of requirements
consider interactions between each stage makes being placed on production planning and control.
design concurrency important. The representa- These requirements have since been categorised
tion proved to be very helpful and so the idea and are discussed in detail later in this section
was extended to link product design, process under the headings general requirements, stock
design and manufacturing system design with reduction requirements and quantitative,
the three parts of ‘system design’ namely: ‘PPC behavioural and administrative activities.
12. 192 Bonney – Reflections on Production Planning and Control (PPC)
Company Market
Requirements Revisions
and Philosophy
Proposed
Determine PPC Structure
PPC Structure
A4.1 Revisions
Determine
System Proposed System
Parameters Specification
Revisions
A4.2
Evaluate
Proposal Proposed Prototype
A4.3
Pilot
Prototype
Proposed
A4.4 PPC System
Figure 2 – Iterative design of the PPC system.
The general requirements being placed on but there is still little experience in bringing
PPC include the need to identify performance them seamlessly together. Actions that improve
measures and to improve performance. The performance can include formalising the adminis-
requirements include the need to improve trative procedures, obtaining better understand-
schedule conformance, provide systems with ing using quantitative studies and working better
greater flexibility and provide greater return on with people. Quantitatively the need is to
capital employed. Inventory planning require- investigate and model the relationships between
ments include the need to provide better service. the variables that affect the performance of PPC
This is usually interpreted as the need to shorten systems so as to be able to choose and then
manufacture lead times, reduce work in progress operate a PPC system that matches the variables
and reduce stocks in the logistics chain but and improves the performance of the company.
occasionally it may require relocating stocks and So far researchers have given greater weight to
sometimes increasing stocks. A variety of quantitative studies based on mathematical
methods may be used within a company analysis and simulation whereas companies have
including using lean production methods, using given greater weight to administrative changes.
smaller batches and reducing safety stocks, There has been insufficient investigation into
managing inventories better by using appropriate improving understanding of human behaviour.
inventory control systems that also provide Behaviourally the need is to understand better
accurate and up to date inventory records and the needs of the community that is the company.
associated control data. Together the items on The people operating the system need to be
this list help production and inventory planning involved. They also need to be provided with
and control to be more responsive. However the appropriate information so that they may use
exact balance and sequence of implementation their local knowledge. The requirements for this
will vary from company to company. are essentially what is called good management.
Co-ordination of multi-level multi-product Behavioural effectiveness is closely linked with
planning and control systems is particularly the administrative needs discussed in the next
difficult. Many of the tools required now exist paragraph. The actions required include:
13. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 193
Market
Demand System Performance
Master MPS
Scheduling
a1
Performance
Purchase Requirements
MRP
Manufacturing Orders
a2
Purchase
Manufacturing Scheduling Purchase Schedule
Scheduling a4
a3
Manufacture Schedule
Figure 3 – A planning hierarchy for an MRP system.
• Educating people so that they know why data including product structures and the
things are being done; operations and the associated processes that are
• Training people so that they feel ownership used to produce the parts and products. It also
of the system and can operate the proposed requires the maintenance of accurate and up to
new system with confidence; date stock records. Technically, it is now
• Involving people and changing things in a possible to collect data relating to all transac-
way that motivates; tions virtually instantaneously. However, if all
• Doing the things that are required e.g. for data were presented to a manager, the volume
legislative reasons to meet Health and Safety would make it impossible to look at let alone
requirements, employment law, etc. in addi- interpret. It is therefore essential to convert this
tion to the actions needed to control produc- raw data into information that is understandable
tion; and usable. This is a requirement of the
• Ensuring that the results are consistent with management information system. Typically this
the needs and wishes of the people operating is achieved by filtering and summarising the raw
and affected by the PPC systems; data and then presenting it, possibly in the form
• Monitoring and modifying systems so that of exception reports e.g. items late, items outside
they are maintained and improved. predefined control limits, etc. Different levels of
aggregation are likely to be appropriate to
Administratively the aim is to produce good different levels of management. The information
structures and systems that co-ordinate manufac- that is presented could conveniently be consid-
turing across the extended enterprise logistics ered to be part of a decision support system
chain. In general, managers intuitively see the (DSS). The information requires to be personal-
need to formalise systems, to control data better ised to meet the needs of individuals, each of
and to simplify systems. Providing realistic whom is different. Above all there is a need to
operational schedules with regard to lead times provide each manager with information
and capacity will reduce uncertainty and reduce appropriate to his responsibilities. However
the need for rescheduling. To operate a information is also motivating and the conse-
quantitatively based PPC system requires quences should be consistent with the perform-
maintaining accurate, up-to-date records of basic ance measures and needs of the company. All of
14. 194 Bonney – Reflections on Production Planning and Control (PPC)
the above has to be done within the culture of changes as part of the investment appraisal
the company. Specific steps include: procedures. Despite these legitimate criticisms
• Identifying the person responsible for each and weaknesses, the issues partially addressed
particular function and its component proce- by MRP have not gone away. A method with
dures. In particular the persons responsible many of the features of MRP is essential to
for providing and maintaining data need to be enable organisations to co-ordinate the availabil-
identified; ity of parts required for the assembly of complex
• Defining the Human Computer Interface products made on an intermittent basis.
(HCI) for each function; Production still needs to be planned and the
• Organising and formalising procedures for activities still need co-ordinating. However, it is
performing each function; important that the methods used should be
• Simplifying the PPC organisation and improved.
procedures. There is also a need to improve operational
Another need is to improve planning proce- control on the shop floor so that the system can
dures. It is particularly important to address the respond to changing operational conditions
issues of determining Master Schedules so that arising from the situation of uncertainty in the
the performance can be potentially the best production environment and the market. One
whatever the constraints. It is particularly way to do this is to incorporate ideas from lean
difficult to plan and control small batch production, particularly those based on local
repetitive production of multi-level, multi-item visual control, into the manufacturing environ-
complex products with variable demand and ments where MRP is currently used. For
long lead times. Despite being unfashionable and example, introducing local control into a push
having many features with which there has been system offers the possibility of combining the
much dissatisfaction, Material Requirements organisational benefits of pull systems with the
Planning (MRP), a practical method developed central planning and information system benefits
within industry, continues to be the method most of current systems. Another requirement is to
frequently used for this purpose. As briefly introduce better shop scheduling.
discussed in Section 3, dissatisfaction with MRP
arises from a number of causes that include 8. Some Supporting Ideas
inherent contradictions in the system related to
lead times and capacity, the need to maintain
highly accurate stock records and difficulties
associated with rescheduling. Further, despite
T he aim is to produce a research agenda that
satisfies the ‘consequential Production
Planning and Control needs’ listed in Section 7.
obvious links with design, process planning and First some ideas that provide a basis for
shop scheduling, most companies require their conceptualising some aspects of the proposed
planning and control systems to live with the research agenda are discussed. They are ‘a
consequences of changed product structures, framework for production control’ and ‘control
obsolescence and configuration control rather theory analysis’.
than becoming part of the multi-variate decision
process. Even investment in new machines and 8.1 A Framework for Production Control
processes that affect capacity and manufacturing
methods is usually decided independently of A framework for studying production plan-
the planning procedures even though lead times ning and control was described in BONNEY &
and potential output will be affected. Therefore, HEAD (1993) and BONNEY et al. (1999b). The
there is a need to develop systems that examine framework may be represented schematically by
the planning consequences of such proposed Figure 4.
15. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 195
Plan or
MIS
Simulate
Action
Plan Information
Disaggregation Aggregation
Figure 4 – A production planning and control framework.
The framework was proposed to represent the • provide a way of examining PPC systems at
structural commonality of production control the structural level;
systems. The characteristics of the framework are • provide a method of classifying PPC systems;
a planning hierarchy, a hierarchy of simulations, • allow theory, simulation and practice to be
a reporting or management information system represented in a common way and so allow
hierarchy and a set of feedback loops. Theory, the methods of one approach to be more
practice and simulation can be represented easily appreciated and possibly used by
within this common structure. The planning another;
hierarchy dis-aggregates high level plans into • provide a basis for allowing one system to
more detailed plans. For example, MRP converts evolve into another.
a Master Production Schedule into the require- Each of the boxes in the schematic frame-
ments for parts and materials needed to make the work represents the transformation of specific
products while also taking account of perform- inputs into specific outputs and each of the lines
ance against previous plans. These requirements connecting the boxes is an information flow, the
may then be converted into schedules for direction of which is shown by an arrow. The
purchasing and manufacture as shown in Figure structure of the framework is consistent with the
3 and it is on the basis of these schedules or IDEF structured analysis representation, input-
plans that action is taken and the results recorded output models and control theory models. For
as transactions. The simulation hierarchy is part example, in a control theory model the boxes
of the planning process and provides information shown in the left hand side of Figure 4 could be
on the likely performance of the plans. The man- transfer functions representing the different parts
agement information system (MIS) hierarchy of the planning system and the boxes in the right
processes and aggregates the transaction data to hand side could be transfer functions represent-
produce performance reports. These perform- ing the processing performed by the information
ance reports are used as feedback information to system. The level of detail that is included is
indicate to the planning functions that plans chosen to correspond with how the framework is
should be modified or other remedial action going to be used.
taken if required. A framework model clarifies the relationships
The reasons for creating the framework were to: between the variables in production control
16. 196 Bonney – Reflections on Production Planning and Control (PPC)
systems, indicates possible omissions or defi- Many systems have been represented using
ciencies in a particular system and provides a the framework and the associated UNISON
way of describing an evolutionary path from one software. These include the modelling of FMS,
system to another. Potentially the parameters of the use of the software as a tool for enterprise
the framework are the basis for a classification integration described in BONNEY et al. (1992),
of production control. Important parameters and a comparison of the performance of push
include the number of levels in the hierarchy, the and pull manufacturing systems described in
degree of parallelism in the system and the BONNEY et al. (1999a). Another problem inves-
relation between the time based variables such as tigated was inventory planning in the context of
lead times, planning horizons, planning period enterprise integration described in BONNEY
and rescheduling frequency. The framework et al. (1996). The model that was used included
structure may be used to develop theoretical and the conversion of a master production schedule
simulation models. However, simply to make into a detailed shop schedule. Although
these assertions is unconvincing because the managers in conjunction with software achieve
whole approach is abstract. There is a need to be this moderately successfully in most companies,
able to illustrate and test the structure and show it is surprisingly difficult to formalise the logical
how it can be used. The UNISON software rules. Indeed there appears to have been little
(BONNEY et al., 1992) based on Petri-nets was insightful analysis of the problem of ensuring
created at the University of Nottingham to allow that detailed plans are consistent with aggregate
this to be done. The UNISON software allows requirements. In principle this may be achieved
one to simulate proposed structures. Hence the by using mathematical programming methods
framework can represent dynamic systems and but, in practice, the usual reason for detailing
within these derive and evaluate the effects of plans is to add further information often
changing values of the variables. Petri-nets are including the knowledge of the human planner.
well described in several text books e.g. This is an example of the need for human
DICESARE et al. (1993) and PROTH et al. involvement outlined in Section 4. Also working
(1996). UNISON can represent the generalised on a finer time scale introduces further con-
framework and data converts this to represent a straints. Appreciating this practical difficulty
specific system structure. This representation was a direct consequence of taking a structural
can then be used to create a simulation model view of different control systems. Likewise it is
that can be used to predict the system perform- difficult to ensure that equivalent bases are
ance in response to given inputs. If the perform- chosen when comparing the performance of
ance of a specific structure is poor and there are different production control systems such as a
structural omissions in the system representation kanban-based pull system and a repetitive batch
compared with the generalised structure e.g. a push system.
feedback loop is not present, then the effect of
‘correcting’ the structure could be considered. 8.2 Control Modelling
The values of the parameters such as those
mentioned above define a specific system Transform methods have been used for
structure. Changing the values of the parameters analysing production and inventory control
in the model of the generalised framework could systems since the seminal work by SIMON
thus represent a move from one system to (1952). These methods provide an understanding
another. If UNISON is used to represent the of the influence that variables have on system
transition from one system to another this allows stability and system performance. Until recent
the system performance of each system to be years however the use of transform methods for
assessed individually and during the transition. studying PPC has been intermittent rather than
17. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 197
systematic. Even now the methods are seldom perform effectively. Other studies examined how
used by industry as part of their PPC system the number of levels in the product structures,
design procedures. The next few paragraphs determined during the process of design for
outline some results obtained by applying manufacture, affect system performance.
control theory to PPC. For some years now, a group under the
If a block diagram representing a system can supervision of Grubbstrom at the Linkoping
be drawn and the transforms and transfer institute of technology in Sweden has been using
functions are known, then an analysis can be Laplace transform methods, input-output
performed. In particular, if the transfer functions analysis and net present value methods to study
that represent the boxes in the framework are the performance of MRP. Laplace transforms are
known then we immediately have a model of the a general method for analysing linear control
specific system in framework format. Popple- systems or linear differential equations. Laplace
well used z-transform methods to analyse the transforms include z-transforms as a subset. This
response of re-order cycle systems and MRP work, a good summary of which can be found in
systems to standard inputs such as an impulse, a GRUBBSTROM & TANG (2000) has extended
step or a sinusoid. A generalised model of the the ideas to stochastic studies and discrete
different information flows of these two multi- decision rules. This has considerably enhanced
product, multi-level planning and control our understanding of the performance of MRP
systems was constructed and the system transfer systems including the problem of backlogs.
functions derived. The stability and performance Other relevant studies using control theory are
of these systems in response to different demand the use of modern control theory by O’Grady,
patterns was examined both analytically using z- who among other models developed a generali-
transforms and by simulation. Some of these sation of the HMMS linear decision rule (HOLT
analyses were reported in POPPLEWELL & et al., 1960) applicable to multi-product, multi-
BONNEY (1987) and in BONNEY & period systems e.g. see O’GRADY & BONNEY
POPPLEWELL (1989). Matoug subsequently (1985). TOWILL (1992a) reviewed control
extended this work to study the effect that theory applications in manufacture and in
misinformation, e.g. inaccurate stock records TOWILL (1992b) described work at Cardiff that
and the effect of delays in providing information linked Industrial Dynamics type simulations
to these systems, had on system performance. with control system analyses. WIKNER (1994)
These studies were partially reported in describes work that links the Linkoping and
BONNEY, POPPLEWELL & MATOUG Cardiff studies.
(1994). Again the system transfer functions were The importance of control modelling is that it
derived and analyses made. Initially it was has laid the basis for a deeper understanding of
assumed that the data such as product structure the reasons why PPC systems behave as they do.
information was available and correct, that stock Converting this understanding into effective
levels were correct, that information was operational rules promises great improvement in
available instantly and everything behaved as the future performance of PPC systems.
planned including manufacturing lead times and
the planned delivery policy. Then various 9. An Outline of the Actions Required by PPC
imperfections were introduced e.g. to analyse the
effect that misinformation and delays have on
system performance within the re-order cycle
and MRP systems. A particularly interesting
T able 1 is used to illustrate the actions that
should be taken by PPC to respond to the
needs identified in Section 7. The rows show
result was the confirmation that high stock some of the major areas of concern e.g. Row 2
accuracy is essential if a MRP system is to shows the ‘General Requirements’ suggested in
18. 198 Bonney – Reflections on Production Planning and Control (PPC)
Section 7, namely that planning needs to be comments are shown in Italics that relate to
more frequent, that planning periods should be Table 1 and to other relevant sections of the
shorter and that there should be better dynamic paper.
response to changes. The columns of the matrix There could be large amounts of data linked
show the major steps involved. As an example, into decision support systems with appropriate
Row 1 is the overall task of introducing a and possibly changing performance measures.
product into production. The steps are product There is a need for appropriate (hopefully
design, workplace design (called Manufacturing reduced) stock levels and for systems to be more
systems design in Section 7 and Figure 1), flexible and easier to reconfigure than at present.
system design and then ‘implement and operate’. It is particularly important to understand how
The individual cells in the table represent production parameters, particularly planning
actions that should be taken or research that is horizons, planning periods and planning
needed. Investigating what to enter into each cell frequency, affect PPC performance. This better
formalises the steps of the total system design in understanding will improve confidence that
a way that is consistent with a concurrent planned actions will produce the required results.
engineering approach. When there is uncertainty In addition to this formal (theoretical) under-
about what actions to take, it will be difficult to standing there is a need for these methods to be
define the cell entry and this should prompt an converted into tested operational rules and for
information search. When the information is the PPC staff to be sufficiently well educated
non-existent or unsatisfactory, it points to the and trained that they can operate comfortably
need for research. A finer division of the and effectively with the new paradigms. Some of
columns or cells may be used to provide a the DSS requirements have been summarised in
clearer focus. For example, sub-dividing the rows 6, 7 and 8 of Table 1 labelled ‘data source’,
heading PPC System (part of system design) into ‘data files’ and ‘filter aggregate and compare’.
aggregate planning, master scheduling and The implications of the matrix are still being
operational planning and control can help to actively considered.
clarify the PPC research problems. Then, the
tasks of selecting the best aggregate plan, the Introducing a product into production (Top
best MPS, the best shop schedule and appropri- Row)
ate feedback are entered in Row 3 because they To understand the problems that arise when
determine the appropriate stock levels. To introducing a product, first choose a method e.g.
identify the research requirements one could ask structured analysis that can represent the steps of
whether it is known how to choose the best the product introduction process, then create the
solution even under restricted conditions. One representation and then examine the implications
could then extend the questioning to consider of the representation. Some of these individual
quantitative, behavioural and administrative steps are now expressed in greater detail.
headings. One could further consider whether a • Choose an appropriate representation for the
proposed solution would still match the major steps of the product introduction
requirements if all of the major steps were process. (Top Row)
considered (the top row) and whether it would • Represent the generalised product introduc-
have knock on effects, possibly adverse, on tion process using a ‘concurrent engineering’
items in the left-hand column. The steps representation. The steps include design for
involved in ‘Introducing a product into produc- function, design for manufacture, manufac-
tion’ row and the ‘PPC system’ columns are turing system design, PPC system design and
described in greater detail below to clarify the business system design. (See Figures 1, 2 and
approach further. For some of these steps 3 of section 7)
19. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 199
Table 1 – Introducing a product into production. Some steps in system development,
implementation and operation.
20. 200 Bonney – Reflections on Production Planning and Control (PPC)
• Use the representation to examine the different organisations working with different
implications of proposed products, process or planning periods such as days, weeks, 4
manufacturing system design changes. (Sec- weeks and calendar months. More important
tions 5 and 6) is that the dynamics of the interactions and
• Develop a systems framework for incorporat- delays become more noticeable over the
ing theoretical developments into future supply chain and minor fluctuations in de-
systems. (Figure 4) mand may get greatly amplified. (Analyse as
• Understand the effects of changing the in Section 8.2)
system structure. (Analysis of models based • There will be shorter time to change models.
on Figure 4) The effects of introducing a new model
• Understand the effects of changing the rest of spread throughout the chain and there are
the manufacturing system. risks of obsolescence when the model stops.
• Determine the response of a proposed system • The financial implications of delays in
to changes in the market. (Different inputs payment and the cash in the supply chain and
into Figure 4) the costs of stock may be harder to identify
• Determine the relevant system variables to particularly when there is rapid movement
study and determine how they influence between organisations e.g. with JIT much of
system performance. (Mathematical analysis, the stock may be in transit.
Sensitivity analysis and simulation) • Insights are needed into how E-commerce
• Investigate the effect of structural features might enhance supply chain co-ordination
such as the number of levels in the hierarchy and planning.
on manufacturing system performance. (A
detail of Figure 4) Determine and implement the operational
• Implement and operate the system. system.
• evaluate methods for co-ordinating the
‘PPC system’ design. (This ‘Design System’ availability of parts and tools for complex
column is a detail of ‘Introducing a product into assembly work with demand;
production’) • determine the implementation procedures;
• match a company’s PPC system response to • develop a prototype operational system repre-
the products and the environment (technology sentative of the next generation of systems
and market) in which the company operates for planning and controlling manufacture;
(Sections 5, 6 and 7); • implement the system;
• determine the factors, including the range and • operate the system and maintain control;
complexity of the products and the variability
of the market demands, that influence the Choose an appropriate Decision Support
choice of system; System to provide responsive planning and
• choose the levels of aggregation for planning; control.
• choose the values of time-based variables e.g. • choose the levels of aggregation for reporting;
planning horizons, planning frequency and • choose performance measures;
rescheduling frequency; • determine the level of discretion that should
• check that the response is stable (See section be given to individual managers;
8.2). • determine how to personalise information for
users;
Investigate supply chain problems. (This is the • introduce local control;
‘External System’ column of ‘Design System’) • use methods of visual control to improve load
• The supply chain may need to co-ordinate control and the operational performance,
21. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 201
controllability and flexibility of MRP-based • choose the supply chain;
installations; • choose the horizontal planning sub-divisions
• choose the MIS. in the company e.g. between different shops,
different products, different cells, etc.;
Possible topics to consider under the • choose the planning hierarchy;
quantitative and administrative headings now • create systems that encourage Concurrent
follow. Additionally, and not discussed further, Engineering;
there is a need to implement and improve • retain integration as systems are changed;
understanding of behavioural (Human-centred) • choose an appropriate system structure.
systems e.g. scheduling.
10. Understanding and Administering Changes
Develop quantitative modelling to: to PPC Systems – An Example
• understand the relation between variables that
•
affect the performance of PPC systems;
evaluate theoretical methods for planning and
co-ordinating multi-stage, multi-level produc-
M ulti-product, multi-level production is
very common, yet current methods for
planning and controlling multi-product, multi-
tion; level production are not satisfactory. The most
• develop tools to help determine the Master commonly used system is MRP but, as indicated
Production Schedule; in Sections 3 and 7, MRP does not completely
• choose a Master schedule; meet users needs. The main needs include
• control the aggregate workload; determining the MPS, overcoming the contradic-
• use simulation models to evaluate the appro- tions related to lead times and capacity,
priateness of push and pull planning systems; changing the MRP system representation from a
• develop models of the extended enterprise; static, deterministic representation to a dynamic
• ensure that a detailed shop schedule is representation, maintain highly accurate stock
consistent with the master schedule; records, reschedule more easily, link MRP with
• ensure that equivalent and comparable bases design, link MRP with investment appraisal
are chosen when comparing the performance models, etc. The extensive needs, the ongoing
of different production control systems. This difficulties and the importance of the appropriate
is an investigation of the effect of different planning, suggest that PPC of multi-product,
input-output transformations; multi-level production is suitable to illustrate
• understand and overcome system nervousness; how systems may respond to the kind of changes
• determine the dynamic planning characteris- described in this paper. The rest of this section
tics; discusses how to use the ideas discussed earlier
• estimate lead times; in the paper to improve performance.
• choose the system parameters including time Changing PPC systems takes time. There is
related parameters e.g. planning horizons and therefore a choice whether to improve the PPC
planning frequencies and levels of aggrega- systems within the existing structure or whether
tion for planning and control. to consider changing the structure. Whichever is
chosen, it is useful to consider administrative,
Provide support for administrative systems. behavioural and quantitative actions that are
Actions include: needed to improve PPC performance. Adminis-
• consider steps such as those listed in ‘Organ- tratively we want the system to work better.
isational and Procedural Actions’ in Section 4; Preferably the administrative steps should
• choose an appropriate company structure for include obvious simplifications, being consis-
the product and market; tent, improving the quality of data, etc, as
22. 202 Bonney – Reflections on Production Planning and Control (PPC)
discussed earlier. Behaviourally we want people tions of PPC can also represent logistics
to be involved and to operate the system better planning problems using the same modelling
and this requires acceptance and ownership of structure. Hence the same theory and structure
the PPC system by the staff. The quantitative may be used to model the complete supply
actions need to be based on sound theory. chain, production with disassembly, recycling
For reasons of comprehension, a hierarchical flows, reverse logistics as well as disposal, repair
approach to PPC is used. In MRP, the hierarchy problems and design modifications occurring
is illustrated in Figure 3. The three steps are during production. Quantitatively, Laplace
usually: producing the MPS, secondly the transforms, z-transforms and modern control
procedural breakdown into requirements of parts theory and the other theoretical representations
and materials and, thirdly, action plans interpret- of MRP based systems discussed in Section 8.2,
ing the requirements into make (shop schedul- have improved our understanding of multi-stage,
ing) and buy (purchase scheduling) decisions. multi-level production-inventory. They offer the
Ideally one wants first to improve Master possibility of changing from static to dynamic
Production Scheduling, an area that currently is representations and have improved understand-
more of an art than a science. This can be ing sufficiently that a research agenda can be set
achieved by choosing feasible solutions and then to improve system performance. First, Master
move some way towards optimised programmes. Production Schedules, product structures, lot
To make this more manageable these plans may sizes, lead times, safety stocks, etc., and the
be set into an aggregate long term planning process of MRP can be represented using matrix
context and the schedules may be split into plans equations. Requirements planning problems can
for different shops, suppliers, etc. In the context be represented in a consistent way whatever their
of the framework shown in Figure 4 these are size particularly with respect to the number of
parallel actions in the disaggregation procedures. levels in the product structure. Matrix methods
For discussion purposes improvements are are essential for the mathematical formulation of
considered within the context of the current these problems and, although the data is sparse,
system structure e.g. continuing to use MRP and matrix solutions are no longer a problem with
maintaining the current levels of hierarchy large storage capacity cheaply available with all
which are assumed to be MPS, requirements modern computers. Secondly, by generalising
planning and scheduling purchases and the the analytical and modelling possibilities of
shops. The main improvements under the input-output analysis and control theoretic
administrative, behavioural and quantitative methods, multi-stage, multi-level production-
action headings are summarised in Table 2. inventory systems can be more fully understood,
First we want a theory that can explain clear criteria set for improving their dynamic
things. Secondly, we want to combine the theory performance and values of the major parameters
with simulation. Thirdly, we want to turn the determined. The theory helps to derive feasible
theory into operational actions. One possible production plans and improve our understanding
way to proceed is to obtain better understanding of how such parameters affect the dynamic
by analysing MRP systems using conventional performance of an actual or proposed manufac-
control theory analysis (transform methods) and turing system. The theory may be used for
the techniques of input-output analysis. Together evaluations from feasibility studies to system
they offer the opportunity to greatly improve the optimisation. It provides shortcuts for evaluating
performance of MRP systems and, combined the impact of changes in the market and
with simulation, they provide the basis for technical environment at the shop-floor level. It
designing systems that have the required thereby offers additional flexibility to companies
performance. These new theoretical representa- with consequences that should be better system
23. GESTÃO & PRODUÇÃO v.7, n.3, p.181-207, dez. 2000 203
Table 2 – Summary of actions to improve PPC System.
Administrative actions Behavioural actions Quantitative actions
Agree actions on make, Improve MPS realism
Improve MPS realism
MPS order acceptance and
using resource planning Simulate proposed MPS
financial control
Input
Improve and maintain
Use new methods to
data e.g. product Encourage Ownership
Data derive data
structures and stock of data
automatically
records
Training and team Improve requirements
Formalise the planning working planning, resource
Planning steps Improve shop
Work better with loading and shop
scheduling
suppliers scheduling methods
System
Improve feedback Use local visual load
Model planning and
Control control and combine
Improve rescheduling control as in 8.2
with best features of JIT
Improve MIS
Output Plans etc
and documentation
responsiveness, better schedule adherence and offering suggestions and making evaluations on
better delivery performance. Inventory levels new schedules in a proactive manner.
will be reduced. There are still practical Developments in computing power enable the
problems and limitations of the theory e.g. an inversion of large-scale matrices to be performed
outstanding problem is how to deal with in real time. This should allow the theoretical
rescheduling. However, the potential economic developments to be turned into robust and
gain from using these methods is great. responsive practical planning methods. These
An input/output model of an enterprise can be planning methods can be used to determine
decomposed into a set of local models, with master schedules and perform capacity planning
overall constraints to ensure that total system in a dynamic situation, determine when and how
performance is maintained. The display, retrieval best to reschedule, vary lead times according to
and communication abilities provided by modern the workload and decide how to deal with
information technology can then be utilised in uncertainties, safety stock and safety lead times.
combination with local modelling to provide There is a need for prototype software that can
improved and visual methods for local control. perform the above tasks, so that the methods
As well as local within the manufacturing unit, may be tested in companies and so that the
this becomes important when including the performance of the proposed system may be
upstream and downstream activities of the simulated and operational systems tested. These
supply chain. The development of e-commerce developments also need to be tested operation-
raises the possibility of an integrated system ally in combination with local visual control.
with suppliers and customers entering data, This will allow managers and schedulers to use