MetaMorph: An adaptive multi-agent architecture for advanced manufacturing systems.

摘要

Global competition and rapidly changing customer requirements are forcing major changes in the production styles and configuration of manufacturing organizations. Traditional centralized and sequential manufacturing planning, scheduling, and control mechanisms are insufficiently flexible to respond to changing production styles (for example, reduced batch sizes and increased product mix) and highly dynamic variations in product requirements. New architectures for the next generation of manufacturing systems will be agent-based systems with novel approaches for dynamically creating and managing agent communities in widely distributed and ever-changing manufacturing environments.; In this dissertation research, a prototype architecture for integrating extended-enterprise issues into open, scalable, reconfigurable, intelligent, and learning manufacturing systems is presented. The prototype implementation of this architecture includes coordination protocols for intelligent multi-agent manufacturing systems based on the creation and destruction of virtual dynamic agent organizations. A new paradigm for the construction of virtual control architectures is therefore a specific contribution of this dissertation. The prototype architecture has been called MetaMorph.; A multi-agent virtual manufacturing system has also been implemented (as a simulation), using the MetaMorph architecture for extended-enterprise requirements. Three novel features are incorporated in this multi-agent virtual manufacturing system: integration of planning and scheduling, plan refinement, and learning. Plan refinement is an elegant feature added to the system to enhance accuracy using simulated real-world information in each manufacturing agent domain. Learning is a highly valued feature for intelligent agent systems--especially for intelligent systems in practical industrial applications.