Optimization of business game decisions
Abstract
Business games place people into a virtual economic environment in which competitive decisions have to be taken. In this context, an objective is to design intelligent agents that may take decision in place of the players. Therefore, the global problem should be modeled, taken into account all the constraints and the specific way to estimate the quality of the decisions. Then a hybrid optimization approach may be developed, combining a local search with exact optimizations of subparts of the problem. Experiments evaluate this approach with real played sessions.