A Layered, Multi-Agent System for Intelligent Control
Takuya Ishioka and Morikazu Takegaki
Industrial Electronics & Systems Lab., Mitsubishi Electric Corp., 8-1-1 Tsukaguchi Honcho, Amagasaki, Hyogo 661, Japan
We propose a layered, multi-agent model as a method for establishing an intelligent control system. The model handles each task, which constitutes control processing, as a group of agents operating autonomously, and organizes the agents in a layered structure from the viewpoint of real-time constraints. This enables real-time control processing and knowledge information processing simultaneously. We will describe a method for building the layered multiagent model using a concurrent rule set. The concurrent rule set can be processed by extending the expert shell and can realize the layered agent model in various computer environments through the use of the shell.