Reducing communication load on contract net by case-based reasoning-eavesdropping for utilizing message leakage

This paper describes communication load reduction on task negotiation with contract net protocol (CNP) for multiple autonomous mobile robots. For controlling multiple robots, CNP is useful, but the broadcast of task announcement messages on CNP tends to consume much communication load. In order to overcome this problem, the authors have developed a system called LEMMING which learns proper addresses for the task announcement messages with case-based reasoning. However, the learning method used in LEMMING sometimes caused inefficient task execution. In this paper, we propose an extension of LEMMING with message interception which enables the system to execute tasks more efficiently by eavesdropping on leaked message.