An adaptive, maintainable, extensible process agent

Abstract

Intelligent agents should be autonomous, cooperative and adaptive. An intelligent multi-agent architecture for business process applications also supports maintenance and extensibility. Agent adaptivity is based on observations of the agent's environment, of the results of previous agent decisions, and of estimated reasons for those results being what they were. Maintainability enables agents to do things that are both correct and politically astute. Extensibility enables the agents to develop ways for doing new things. Agent autonomy and cooperation are achieved by building the architecture on a three-layer BDI, hybrid, multi-agent architecture.