TOWARDS A GENERAL THEORY OF SYSTEMS. MODELS OF SUBSYSTEMS.

Abstract : The classification of Operations Research models into eight types is arbitrary and largely the result of historical accident. The models are largely designed for use with systems involving only one decision maker. The research described in this report began with the idea that insight into the behavior of complex systems could be obtained by a search for more general models involving two or more of the eight model types. A plausible method of optimizing the behavior of complex systems is an iterative procedure in which each subsystem is optimized in turn. It is shown by example that this method is unreliable in that even when the iteration converges, it does not necessarily converge to the system optimum. Consequently, it is necessary to construct models emphasizes a need for mathematical development, particularly in the solution of partial difference equations. The approach through a complete model, however, might misrepresent the problems of any particular decision maker since this approach implies knowledge of the strategies of each decision maker in the system. It is argued that the behavior of complex systems should be investigated by simulation since the information presented to each decision maker can be given in the form in which he would receive it in real systems. (Author)