Norm-based Enterprise Agent Intelligence Design

Abstract

In the fierce market competition, enterprises constantly evolve themselves according to market requirements to better survive and develop. This paper establishes a multi-agent enterprise model, decomposes the complex enterprise system into multiple Agent entities to form a multi-agent system (MAS), and jointly solves the overall goals of the system by multi-agents. Using genetic algorithms and learning classifiers, the concept of norm in organization semiotics was introduced. Then enterprise agents was objectively controlled and constrained by the norm based on the complex semantics of the specification. On this basis, a parallel, rule-based, enterprise intelligence was developed, which can be automatically updated with rules and reflects the subjective initiative of agents under the environment. Finally, a case study was carried out on the Swarm simulation platform. The simulation results show that the design of the enterprise agent intelligence provides support for selection of the dynamic behavior of the enterprise agent.