Subject: Do We Want to Talk about the

Posted by Jay W. Forrester on 2/25/2006
In Reply To:Do We Want to Talk about the Posted by Michael Round on 2/19/2006

Message:

This query about goals leads to an interesting point. As I said in my comment, the local goals at each decision point interact to produce system behavior. We usually model a system, as it is, to understand how the structure and policies (the local goals are part of the policy statements that control individual streams of
decisions) produce the good or bad behavior that is occurring.

If we do not like the actual behavior, and the corresponding model behavior, it means we should do a system redesign. In other words, we have a goal for the system that is not being met by the structure and policies of the operating system. We then try changes in structure and policies in the model to determine what changes to make in the actual system.

There is another way to look at this question. Indeed, people in the system have "global goals" for system performance. For example, a company may want to increase profits or market share. But what
effect do those goals have? Are they operational? Can one trace
specific influences of those goals. Often one can see that there is an effect of the goals, and often the results are entirely counterproductive. People often do not judge correctly the effects that come from assertions about overall goals, For example, the system difficulties may arise because of the dominance of short-term objectives and the failure to meet vague global goals may increase the short-term pressures that are causing the problems.

To summarize, "global goals" have an effect only to the extent that they influence decision making at local points in the system. One does not get improved behavior by wishing for it. Good results come from insightful redesign of structure and/or policies at one or more local points. The intuitive influences of global goals usually produce the wrong results because they act through channels whose operation may give unintended results.

A system dynamics model should start by modeling the system as it is, including the specific effects of global goals, if there are any meaningful global goals. The the model can be used to explore how to change system design, including the effect of vague global goals, to improve behavior.

Follow Ups:

Do We Want to Talk about the - Louis Macovsky 2/27/2006 
Do We Want to Talk about the - Louis Macovsky 2/25/2006 
Do We Want to Talk about the - Louis Macovsky 2/27/2006