Cooperative Multi-Agent Learning: The State of the Art
Autonomous Agents and Multi-Agent Systems
Focus of control through goal relationships
IJCAI'89 Proceedings of the 11th international joint conference on Artificial intelligence - Volume 1
| Hi-index | 0.00 |
COOPERATIVE DISTRIBUTED PROBLEM SOLVING NETWORKS ARE DISTRIBUTED NETWORKS OF SEMI-AUTONOMOUS PROCESSING NODES THAT WORK TOGETHER TO SOLVE A `SINGLE'' PROBLEM. THE DISTRIBUTED VEHICLE MONITORING TESTBED IS A FLEX- IBLE AND FULLY-INSTRUMENTED RESEARCH TOOL FOR EMPIRICALLY EVALUATING ALTER- NATIVE DESIGNS FOR THESE NETWORKS. THE TESTBED SIMULATES A CLASS OF DIS- TRIBUTED KNOWLEDGE-BASED PROBLEM SOLVING SYSTEMS OPERATING ON AN ABSTRACTED VERSION OF A VEHICLE MONITORING TASK. THERE ARE TWO IMPORTANT ASPECTS TO THE TESTBED. FIRST, IT IMPLEMENTS A NOVEL GENERIC ARCHITECTURE FOR DISTRIBUTED PROBLEM SOLVING NETWORKS THAT EXPLOITS THE USE OF SOPHISTICATED LOCAL NODE CONTROL AND META-LEVEL CONTROL TO IMPROVE GLOBAL COHERENCE IN NETWORK PROBLEM SOLVING. SECOND, IT SERVES AS AN EXAMPLE OF HOW A TESTBED CAN BE ENGINEERED TO PERMIT THE EMPIRICAL EXPLORATION OF DESIGN ISSUES IN KNOWLEDGE-BASED AI SYSTEMS. THE TESTBED IS CAPABLE OF SIMULATING DIFFERENT DEGREES OF SOPHISTICATION IN PROBLEM SOLVING KNOWLEDGE, OF PROCESSING (SIMULATED) INPUT DATA WITH VARYING AMOUNTS OF DISTRIBUTION AND AMBIGUITY, AND OF TESTING DIFFERENT MECHANISMS FOR CONTROLLING LOCAL AND NETWORK PROBLEM SOLVING (INCLUDING ORGANIZATIONAL STRUCTURING, LOCAL PLANNING, AND PARTIAL GLOBAL PLANNING). NODE CONFIGURATIONS AND COMMUNICATION CHANNEL CHARACTERISTICS CAN ALSO BE INDEPENDENTLY VARIED IN THE SIMULATED NETWORK.