Bdsm-ch3 Modeling Process

Business Dynamics and System Modeling

Chapter 3: Modeling Process Pard Teekasap Southern New Hampshire University

Outline 1.Purpose of modeling 2.The client and the modeler 3.Steps of the modeling process 4.Modeling is iterative 5.Overview of the modeling process

Purpose of modeling • Who are the most important people in the safe operation of an aircraft? • Designer OR Pilot • Too many managers spend far too much time acting as pilots rather than creating an organizational structure consistent with their vision and values

Small or Big Problem • Solutions to small problems yield small rewards • Focus your model on the important issues, on the problems where your work can have lasting benefit, on the problems you care most deeply about

Client and Modeler • Clients = people you must influence for your work to have impact • Purpose is to help the clients solve their problem. If the clients perceive your model does not address their concerns or lose confidence in it, you’ll have little impact

Clients are not always right • Modelers have a responsibility to require their clients to justify their opinions, ground their views in data, and consider new viewpoints • When the clients ask for something you think is unnecessary or misguided, you must work with them to resolve the issue

Steps of the modeling process 1. Problem Articulation (Boundary Selection)

5. Policy Formulation & Evaluation

4. Testing

2. Dynamic Hypothesis

3. Formulation

Iterative with real world Real World

Decisions (Organizational Experiments)

1. Problem Articulation (Boundary Selection)

5. Policy Formulation & Evaluation 4. Testing

Strategy, Structure, Decision Rules

Information Feedback

2. Dynamic Hypothesis

3. Formulation

Mental Models of Real World

Problem articulation • What is the issue the clients are most concerned with? • What problem are they trying to address? • What is the real problem, not just the symptom of difficulty? • What is the purpose of the model?

Can I model the whole world? • Is it useful to have the map as detailed as the territory? • To be useful, model must address a specific problem and simplify • Usefulness of models lies in the fact that they simplify reality, creating a representation of it we can comprehend • The model should be simple enough so the assumptions could be examined

Methods for articulate the problem • References Modes • Time Horizon

Reference Modes • Set of graphs and other descriptive data showing the development of the problem over time • You and the clients must identify the time horizon and define variables and concepts you consider to be important for understanding the problem and designing policies to solve it

Time Horizon • How long should I collect the data? • It should far enough back in history to show how the problem emerged and describe its symptom • How far should I simulate to? • It should far enough into the future to capture the delayed and indirect effects of potential policies

US Energy market is stable 18

Million Barrels/Day

Consumption

12

Imports

6

Alaska

Production, Lower 48 States

0 1986 30

1988

1990

1992

1994

1996

1988

1990

1992

1994

1996

1990 $/bbl

25 20 15 10 5 0 1986

Are you sure it is stable? 18

Million Barrels/Day

Consumption

12

Imports Alaska

6 Production, Lower 48 States

0 1870 50

1890

1910

1930

1950

1970

1990

1890

1910

1930

1950

1970

1990

1990 $/bbl

40 30 20 10 0 1870

Fossil Energy Production

Or it is stable?

-10000

-5000

Iron Age

Year

0 Industrial Revolution

5000 First Oil Shock

So how long should I simulate • A good rule of thumb is to set the time horizon several times as long as the longest time delays in the system, and then some

Formulating a dynamic hypothesis • Develop a theory to account for the problematic behavior • Dynamic because it must provide an explanation of the dynamics characterizing the problem in term of underlying feedback and stock and flow structure of the system • Hypothesis because it’s provisional, subject to revision or abandonment • Goal is to help the client develop an endogenous explanation for the problematic dynamics

Endogenous Explanation • An endogenous theory = the dynamics of a system through the interaction of the variables and agents represented in the model • A exogenous variables theory explains the dynamics of variables in term of other variables whose behavior you’ve assumed • From the beer game, who generates the system behavior

Exogenous is still important • Focusing on endogenous explanations does not mean never include any exogenous variables in the model. But the number of exogenous inputs should be small and carefully scrutinized • If there are any important feedbacks from endogenous elements to that exogenous input, the model should expand and that input must be modeled endogenously

Tools for mapping system structure • • • • •

Model boundary chart Subsystem diagram Causal Loop Diagram (CLD) Stock and flow maps Policy structure diagrams

Model boundary chart Endogenous GNP Consumption Investment Savings Prices Wages Inflation rate Labor force participation Employment Interest rate Energy production

Exogenous Excluded Population Inventories Technological change International trade Tax rates Environmental constraints Energy Policies Non-energy resources Interfuel substitution Distributional equity

Benefits of model boundary chart • Alert the clients to a questionable assumption so they could evaluate what the effect of the missing feedback might be • Without a clear understanding of the boundary and assumptions, models constructed for one purpose are frequently used for another for which they are ill-suited, sometimes producing absurd

Subsystem diagram • Convey information on the boundary and level of aggregation in the model by showing the number and type of different organizations or agents represented • Should not contain too much detail • Multiple subsystem diagrams can be used to convey the hierarchical structure of large models

Subsystem diagram for corporate growth model Sales Effort Product Suitability Delivery Delay Quality Price Delivery of Product Company

Market Payment Orders Mkt. Response to Price Mkt. Response to Quality Mkt. Response to Delivery Delay Mkt. Response to Suitability

CLD, Stock & Flow, Policy structure diagram • CLDs are useful for diagramming the feedback structure of systems • Stock and flow diagrams emphasize their underlying physical structure and track accumulations of material, money, and information as they move through a system • Policy structure diagrams focus on the information cues the modeler assumes decision makers use to govern the rate of flow in the system

Formulating a simulation model • Transfer from conceptual model to formal model • Formalization helps to you to recognize vague concepts and resolve contradictions that went unnoticed or undiscussed during the conceptual phase

Testing • Not only to replicate historical behavior, every variable must correspond to a meaningful concept in the real world • Every equation must be checked for dimensional consistency • Sensitivity of model behavior and policy recommendations must be assess in light of uncertainty in assumptions • Test for extreme conditions

Policy design and evaluation • Policy design is more than just changing the value of parameters. It includes the creation of entire new strategies, structures, and decision rules • The robustness of policies and their sensitivity to uncertainties in model parameters and structures must be assessed, including their performance under a wide range of alternative scenarios