Abductive coordination for logic agents

In this work, we introduce abduction in a (logic) multi-agent environment to achieve a basic coordination scheme among agents. In particular, we define and implement a distributed protocol to coordinate the reasoning of agents. Each agent performs hypothetical reasoning, achieved by means of abduction. To this purpose, each agent is character&d by a knowledge base represented by an abductive logic program. Agents coordinate themselves on the basis of knowledge represented by hypotheses posted into a tuple space. Agents might differ in their knowledge bases, but must agree on BS sumed hypotheses. We define a coordination protocol and tailor it for a basic abductive proof procedure. We also sketch an implementation, under development, obtained by using Java and SICStus Prolog.