Modelling the Structured and Complex Decision Situations

Game theory, modelling and simulation, artificial intelligence, decision making

The paper deals with modelling and simulation of complex intelligent systems where the modelled entities have to make certain decisions in conditions of for example competitive market. We base this paper on our experiences in making models of competitive markets with electricity and other auxiliary services in area of Czech republic and Central Europe. The studied problem is certainly an excellent example of multi--player competition on multi--market with multi--commodity (a player has to decide what commodity is going to offer at which market in competition to other players). In this paper, we are more focused on theoretical analysis of the problem and its effective implementation in simulation models. We transform these situations of strategic decisions of one or more intelligent entities to so called structured and complex decisions. We are going to present a rather general solution to these systems built from elementary decisions with a special emphasis on their effective simulation. We use a special mathematical formalism called Automated Information Net (AIN) to model relations between elementary parts of the AI model. AIN can predict these computing iterations which would not bring new information to simulation of our models. We save some simulation time by this principle. As our background is mostly in area of economics modelling, we oftenly use market and production examples to demonstrate rather theoretical problems. The use of our ideas is hopefully wider.