Reasoning structures for multi-agent meta-programming

Abstract

Meta-programming a multi-agent system is a complex task due to the fact that as the agents build their strategies the environment changes. We propose a method and a language for multi-agent meta-programming based on explanation based learning and we unify these ideas under a formal framework. As an example, we report our experience of use of our meta-programming method on the example of predators for meta-programming multi-agent systems.