Manufacturing systems: Using agents with local intelligence to maximize factory profit

Many of the problems of manufacturing systems, including scheduling, are NP-hard and therefore cannot be solved and optimized in real time as required in the real world. This paper investigates the possibility of controlling manufacturing systems while maximizing the enterprise profit in a stochastic and dynamically varying environment. In this research, we developed a multi-agents system that operates on the factory floor level and directly controls the manufacturing system. The paper examines the proposed multi-agents system, demonstrates its feasibility by simulation under various experimental settings, presents important findings and excellent results, and draws conclusions.