A knowledge representation and modeling agent for programming models in Operations Research

Applying state-space theory in Artificial Intelligence and Agent technique, this paper presents a knowledge representation and a Agent-based modeling system for programming models in Operations Research In this representation, a programming model can be defined as a n-tuple M-p = ( I, G, S-3, Sn-1, S) and the solution procedure of the model has be transformed into a searching operation for optimal path from initial node I to goal node G in its state-space diagram It is convenient to realize the inference procedure: for symbolic knowledge, model integration and automatic modeling. Moreover, the structure of multi-Agent Modeling System for Programming Problems in OR and the principle to form knowledge-based mathematical model has been put forward, too. The inference procedure in modeling has also been explained by means of a modeling example of a investment decision problem It is beneficial for this research to modeling and real-time control of dynamic.