An Approach to Cooperatio: Planning and Communication in a Distributed

COOPERATION BETWEEN INTELLIGENT AGENTS REQUIRES THAT THE AGENTS UNDER- STAND THE CONSEQUENCES OF THEIR ACTIONS AND INTERACTIONS. AGENTS MUST PLAN TO PERFORM BENEFICIAL ACTIONS, AND MUST COMMUNICATE ABOUT THEIR INTENTIONS TO PROMOTE HELPFUL INTERACTIONS AND TO AVOID HARMFUL INTERACTIONS. IN THIS PAPER, I DESCRIBE AN APPROACH FOR INCREASING COHERENT COOPERATION AMONG AGENTS THAT ARE WORKING TOGETHER ON A SINGLE PROBLEM. INITIALLY, A SIMPLE MODEL OF COOPERATION AMONG SELF-INTERESTED AGENTS IS PRESENTED. NEXT, THE METHODOLOGIES FOR ACHIEVING COOPERATION BETWEEN PROBLEM SOLVING AGENTS ARE INVESTIGATED AND THE IMPORTANT CONNECTION BETWEEN DISTRIBUTED PROBLEM SOLV- ING AND DISTRIBUTED PLANNING IS RECOGNIZED. SUBSEQUENTLY, I DISCUSS PRE- LIMINARY RESEARCH INTO IMPROVING COOPERATION IN DISTRIBUTED PROBLEM SOLVING NETWORKS BY INTRODUCING PLANNING AND COMMUNICATION OF META-LEVEL (COORDIN- ATION) INFORMATION. FINALLY, FURTHER RESEARCH IS PROPOSED TO IMPROVE ON THE PLANNING AND COMMUNICATION MECHANISMS SO THAT COOPERATION CAN BE ACHIEVED IN A VARIETY OF PROBLEM SOLVING SITUATIONS.