Anarchic manufacturing: implementing fully distributed control and planning in assembly

Abstract

This paper demonstrates that a distributed control and planning system can fulfil an idealised mixed-model assembly problem and compete with traditional systems. The anarchic manufacturing system is a distributed planning and control system, based on a free market structure, where system elements have decision-making authority and autonomy. Mixed-model assembly is typically managed centrally for production planning and control, using simplification and hierarchical structures to manage complexity. In developing anarchy, inter-job cooperation is implemented to synergise jobs together and fulfil global objectives efficiently. The anarchic system maximises available flexibility, through embracing complexity, and reduces myopic decision making by maximising an agent’s lifetime profitability. Through agent-based simulation experiments, the anarchic system is compared to fixed and flexible centralised systems. The proposed system outperforms traditional systems when the scenario’s structural flexibility allows agile and delayed dynamic decision making. Additionally, the anarchic system managed dynamic bottleneck disruptions as effectively as flexible centralised systems.

Keywords

• assembly
• distributed systems
• planning and control