Physics-Informed Dynamic Cell Layout Planning for Matrix-Structured Manufacturing Workshops under Volatile Environment

Abstract Modern manufacturing enterprises face increasing challenges from frequent production disturbances and intensified demand fluctuations, making traditional manufacturing system layouts inadequate for rapidly changing production environments. As a resilient manufacturing paradigm under Industry 5.0, matrix-structured manufacturing systems (MMS) effectively accommodate multi-variety and variable-batch production demands. However, traditional layout algorithms for such manufacturing systems lack effective dynamic response mechanisms with limited real-time optimization of manufacturing cell configurations under disturbance conditions. To address this issue, this study proposes a dynamic cell layout planning method based on the Consensus-Enhanced Fruit Fly Optimization Algorithm (CE-FOA), which integrates a consensus-driven evolutionary mechanism into the traditional fruit fly optimization framework to improve global search efficiency and convergence stability. By formulating layout planning criteria and developing an adaptive configuration approach, the proposed method enables real-time optimization of cell layouts in volatile production environments. Moreover, a case study has been carried out in the scenario of an MMS workshop for optoelectronic pod (OP) production. Experimental results indicate that the proposed CE-FOA achieves improved solution quality and faster convergence compared with the traditional Fruit Fly Optimization Algorithm (FOA) and Simulated Annealing (SA). Specifically, CE-FOA reduces logistics costs by 4.0% and 4.9% relative to FOA and SA, respectively, while reconfiguration costs are lowered by 4.3% and 2.2%.