Interactive control of the model: a natural companion to animated simulation graphics

• Authors: Andrew R. Gilman
• Venue: WSC '85 Proceedings of the 17th conference on Winter simulation
• Year: 1985

Many recent publications have lauded the value of an animated graphical representation of a simulation model as a boon to communicating model results to decision makers. Recent commercial product developments reflect the importance the simulation community places on graphical representations. This paper suggests that the use of animated graphics is a step in the right direction but leads naturally to another step: interactive control of the model.It is now generally recognized that a decision maker is more comfortable with recommendations based on a simulation model that he has “seen”. The natural tendency for a manager who sees the manifestation of a problem (e.g. an overflowing WIP buffer) in a simulation animation is to immediately respond “what if we …” If the model provides immediate access to his suggested change through interactive facilities, then the flow of ideas from the decision maker is greatly enhanced. The decision maker begins to get a feel for the system's responses. This does not replace the need to perform more controlled experiments to determine long run tendencies of the model. That is a task different from acquiring the confidence of the decision maker.Another benefit of interactive models is explored in this paper. There are times that we would like the manager's input to the model for the model's sake rather than the manager's sake. Many models contain some element of human decision making. The decision function is typically simplified into a series of logical conditions for the sake of coding. An interactive model provides a means for the computer and human to work together to achieve a more realistic model. Rather than going through the trouble of simplifying and coding human decision making, let the real decision maker provide the decision through interactive facilities when the model requires a decision. Let the human do what it does best, and let the computer do what it does best, i.e. repeatedly process the systematic rules of the model.There are several advantages to models that can provide an animated graphical representation of the system being modeled. The modeler often benefits; more important though, the decision making process benefits. The purpose of this paper, however, is not to report claims about graphics and animation. Its purpose is to demonstrate that animated graphics, combined with interactive control of model characteristics, bring benefits that are greater than the sum of these two features considered separately. In particular, this paper will challenge the traditions of batch type models with two valuable uses for interactive model control. The first focuses on the needs of the decision maker whom the modeler seeks to help. The second exploits interactive capabilities as a means to a more realistic, as well as simpler, model.