An integrated differential evolution algorithm for reconfigurable manufacturing systems

Product platforms consider an effective strategy to empower mass customization (MC) in a reconfigurable manufacturing system (RMS) to have the capacity to reconfigure the hardware and manage resources at all organisational and functional levels. This enables quick modification of production functions in reaction to unforeseen changes in the market or regulatory requirements. Although several platform formation and assembly line techniques have emerged to satisfy the increased demand for MC and enable the manufacture of products with high variety, they also increase the system's complexity and make it less cost-efficient as variety increases. This study aims to minimize the complexity of a production system by reducing the number of separate com-ponents in different production processes. An efficient approach that integrates the Median-Joining Phylogenetic Networks (MJPN) classification tool and differential evolution (DE) in one unified framework is developed to do that. The proposed framework uses the concept of the package to group separate components that are often employed jointly to produce particular products and allows the use of the 'all-unit discount policy', which applies discounts to the purchasing costs of a specific quantity of materials/components. Finally, two case studies are applied to validate the proposed MJPN-DE in different dimensions. The experimental results show that applying the proposed package model saves 57.5% of the purchasing costs of components and 41.35% average setup costs compared to MJPN-and non-platform-based models and increases the postponement effectiveness of the ob-tained layout by 41.51% and 74.11% compared with two other layout strategies. In algorithmic comparisons, the results of comparing MJPN-DE with standard DE and genetic algorithm (GA) show an enhanced performance of MJPN-DE where the total production cost is reduced in the small case study by 63.34% and 38.52%, and in the large case study by 16.90% and 11.51%, respectively.