Decentralized Scheduling of Discrete Production Systems with Limited Buffers

Abstract

Modern production systems have become more flexible and agile. For them, a suitable control and planning method is needed. Agent-based systems offer the required characteristics. The contributions of these work are a remodeling of the Shifting Bottleneck heuristic to include limited buffers, the development of a decentralized multi-agent system for production planning, and the adaptation of the heuristic to the agent system. It concludes with evaluations of real-world production systems.