Computer Simulation for Supply-Demand Interaction

Abstract

The most basic result in economics is that, for perfect competition, equilibrium price and quantity are determined by the intersection of aggregate demand and supply curves. How this equilibrium is achieved and the adjustment process have received less attention. The adjustment process may be based on aggregate demand and supply curves only, via the imaginary actions of an auctioneer or a tatonnement process or on modeling of price and excess aggregate demand interaction. Such models are less plausible because they assume agents have perfect information about a unique price at which all transactions are concluded. Alternative assumptions necessitate explicit modeling of individual agents and their search behaviour, which increases greatly the level of complication of the model. Though it is possible to deal with such models analytically (Stigler 1961, Rothschild 1973, Fisher 1973, McAfee and McMillan 1988), special assumptions are needed to obtain desired results. Lesourne (1992) studies many of these models mostly in a descriptive manner in the framework of self-organizing systems. Dynamic models which explicitly include multiple interacting agents contain many details that are difficult to handle without computers. Computer simulation makes it possible to include almost any assumption about individual agents and their interactions and to find out the consequences of these assumptions on the path toward equilibrium. One of the few examples of such simulation is the work by Albin and Foley (1992) which deals with geographically dispersed agents with costs of communication. Applications of computer simulation to other areas of economics can be found in Forrester (1961), Witt (1986) and Orcutt (1990). The current paper describes a computer simulation system that has been organized to enable simulation of interaction using a wide range of assumptions regarding agents, search, and samples, so that the consequences