Agent-Based Evolutionary Model for Knowledge Acquisition in Dynamical Environments

The basic idea of the approach proposed in this paper is to apply multi-agent paradigm in order to enable the integration and co-operation of different knowledge acquisition and representation techniques. The effective operation of learning process is achieved by evolutionary optimization running at the level of agents’ population. In the discussed variant of the model, each agent uses reinforcement learning, and the obtained knowledge is represented as the set of simple decision rules. The approach is illustrated by a particular realization of the system dedicated to the evasive maneuvers problem, together with preliminary experimental results.