P L N 07 B O3 B Glazner Understanding And Modeling

  • 254 views
Uploaded on

 

More in: Education , Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
254
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
4
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. Understanding and g Modeling Interaction Effects in Complex Engineering Enterprises Christopher Glazner PhD Candidate, MIT Engineering Systems Division 2007 LAI Plenary Conference web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 1
  • 2. Agenda g • Overview of the problem • Hypotheses • Enterprise-level interactions • Useful modeling approaches • Next Steps web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 2
  • 3. Motivation • Many interactions across the enterprise can give rise to behavioral complexity • My experience: the Joint Strike Fighter Program • An engineering enterprise tasked to develop, produce and sustain an highly complex fighter that will become the mainstay of US and allied armed forces for 30 years • Evolving threat environment: the system’s requirements will inevitably change over its lifetime • Complex, large-scale interconnected open socio-technical (CLIOS)system • Many stakeholders, partners with differing value propositions • Open environment: encompasses political, regulatory, social domains • An enterprise-level view is needed to understand and design complex engineering systems such as this web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 3
  • 4. Current Approaches to Understanding Enterprises Enterprise architecture frameworks are Enterprise modeling is fundamentally descriptive, not explanatory reductionist, limited scope • Several “views” of an enterprise are • The enterprise is decomposed into identified, such as organization, “domains”(aka views) process, information, and strategy • Each domain is separately modeled • Each view is individually described to answer problems relevant to that according to a framework such as framework, domain Zachman or FEAF • Unfortunately, many problems take • These views are static and describe place across domain boundaries: the current state • IT architecture doesn’t reflect • It has been tacitly assumed that the y • organizational realities Strategic St t i concerns drive partnering di t i sum of these “views” can answer arrangements that affect the design enterprise-level questions process due to security constraints There is a need for an approach to understand the dynamic interactions that effect behavior across the enterprise web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 4
  • 5. “Views”: The Enterprise Architect s Architect's Flashlight •Not linear •Not independent N ti d d t •Numerous interactions Process Enterprise E t i web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 5
  • 6. Key Hypotheses y y • The interactions that occur across the boundaries of the enterprise architecture views can give rise to emergent behavior (operational dynamics) • This operational behavior is best modeled using hybrid, multi-scale modeling techniques; • These modeling efforts can aid enterprise leaders in anticipating operational behavior attributable to the enterprise architecture and help them modify the architecture to suit their needs. web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 6
  • 7. The Idea The Enterprise Frameworks: Strategy Policy Organization •DODAF The Enterprise •Zachman Process Knowledge Information Product/ •FEAF Architecture Service •Etc.. Organization Product/ Enterprise Knowledge Strategy Service Architecture Process Policy Information Model M d l Inputs Outputs web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 7
  • 8. Interactions in an Enterprise Architecture Framework Policy / External Factors Process Organization Products / Strategy A Services Knowledge Information Technology web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 8
  • 9. The Hybrid Modeling Approach y g Organization Process Strategy View View View System Dynamic Model rete Event Model (Long-Term Strategy) (The Organization) del cs ased Mod (Mid-Term Strategy) (Division) Agent-ba m Discr (Teams) (Short-Term Strategy) (Individuals) Sub-model 1 Sub-model 2 Sub-model 3 web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 9
  • 10. The Merits of a Hybrid Modeling Approach • Behavior at the scale of the enterprise is often emergent, arising from the interactions across the views and across scales. • The behaviors within one of the enterprise views at a certain scale often fits neatly within one context, but these behaviors interact with others that do not. There is no one modeling approach that can capture all of these behaviors and their interaction appropriately. (Mingers and Gill, 1997) • Th idea is to use a modeling approach that is best tuned to each The id i t d li h th t i b t t dt h context/EA view (bottom-up, top-down, sequential, event-based, contingent-based, etc) for each sub-model. • By linking together the multiple contexts/sub-models with feedback, complex behavior can emerge within the model, allowing a user to explore ways to affect changes to these behaviors through changes in the architecture. web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 10
  • 11. Modeling Interactions Between Sub-models • The interactions/interdependancies between the enterprise p p architecture views (cross-view interactions) is central to this work. I have informally termed these interactions within the model “hooks.” They fall into two broad categories: • V i bl key enterprise metrics, agent population sizes, Variables: k t i ti t l ti i characteristics, DE process attributes • Events: DE model events, changes in contingency fit, game outcomes • A hook has attributes that have meaning across the many contexts of the sub-models. • Practically, these hooks are pathways that allow models in different contexts to communicate and influence each other. web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 11
  • 12. Modeling Environment g • I am using the program AnyLogic (www.xjtek.com) for my simulation • Benefits: • Environment natively supports system dynamics, agent-based modeling, discrete event modeling, and Bayesian Network d li di d li dB i N k modeling • Flexible, open JAVA based environment makes it possible to insert custom code and models • Supports exportable applet-based animations and optimization of S t t bl l tb d i ti d ti i ti f models based on specified parameters • I may also use OrgCon in addition to model strategic and organizational contingencies within the model. web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 12
  • 13. Next Steps • Finish development of a hybrid model based on LAI’s Lean Enterprise Value Simulation; • Develop a hybrid model of enterprise p y p interactions based on a case study of enterprise change taken from the aerospace industry i d t • Explore methods/approaches to validation and verification of enterprise models web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 13
  • 14. Review • Focus of work: developing a modeling approach to capture complex enterprise behaviors driven by high level interactions across the enterprise • The approach taken involves hybrid modeling (multi- methodology) at multiple scales, specific to a point change in the enterprise. • The key to understanding such behavior is through analyzing the interactions and interdependencies across the many facets of the enterprise. web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 14
  • 15. Questions? web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 15
  • 16. Current Topics for thought… g • How simple can these models be before loosing the desired insight? • How are such complex simulations verified? • When is hybrid modeling the right way to go? • In what context are local and enterprise decisions being made? • What insight do local decision makers have over enterprise issues? What information is and is not available? What incentives drive local behavior? • How can enterprise leaders determine the right levers p g and incentives across the enterprise? • Can causality relationships be inferred? web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 16
  • 17. A Concrete Example: Lean Enterprise Value p Simulation • A simulation of a complex aerospace enterprise • Philosophy draws heavily on LAI research and the recent book Lean Enterprise Value • Content and cases based on LAI member experience • Created and distributed by Hugh McManus ( c a us@a u (hmcmanus@alum.mit.edu) a d t edu) and Eric Rebentisch (erebenti@mit.edu) web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 17
  • 18. A Tool to Teach Enterprise Transformation Skills • An experience in using lean methods to analyze complex systems • Begin working at local levels to gain familiarity with techniques and understand benefits and challenges • Expand to enterprise to explore the issues of coordination, E dt t i t l th i f di ti communication, analysis, and execution • Better understand enterprise change dynamics • Explore alternative implementation strategies and tradeoff benefits • Repeated cycles of lecture, exercises, and implementation reinforce transformation process lessons i f t f ti l web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 18
  • 19. Enterprise Value Flows and Their Metrics • Major t M j enterprise components i t • Flows (physical, knowledge, money) between them • Key enterprise data and sources of inefficiency y p y • Can show enterprise wide (multi-program) flows web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 19
  • 20. Strengths and Weaknesses of the LEV Model • Strengths • Demonstrates emergent behavior from the flo of information and beha ior flow resources across contextual boundaries • Provides a proof of concept from some aspects of my proposed modeling • Simple; completely specified • Weaknesses • Currently does not use an EA Framework as the basis of the model (currently process based) • As a result, there is little need for hybrid modeling web.mit.edu/lean © 2004 Massachusetts Institute of Technology Presenter/date - 20