Scenario-based multi-objective optimization for manufacturing reliability with production routes and available machine types

Abstract

This study introduces a multi-objective reliability model designed to meet the customization needs of production systems by integrating manufacturing routes and machine types within a multilevel framework. A novel encoding approach and evolutionary operations are embedded within the indicator-based Artificial Bee Colony algorithm (ε-MOABC) to search for near-optimal solutions that accommodate diverse production requirements. A series of experiments were conducted to validate the multi-objective manufacturing scenarios. Initially, the model was tested using serial and serial-parallel systems, demonstrating that higher component reliabilities and redundancy levels are closely associated with shared machine types. The experiments further showcased the model’s effectiveness in managing key machine characteristics, such as maintenance, changeover, and workload. In a case study involving a steel forging plant, the model was extended to optimize the reliability of 12 products while minimizing costs and floor space. The findings emphasize the importance of balancing cost savings, machine efficiency, and transportation logistics. This study additionally explores Pareto-optimal solutions across production lines, providing insights into preference selection. Sensitivity analysis was also conducted to validate the model’s robustness. The discussion includes the assumption of constant failure rates and offers managerial implications, providing practical guidance for optimizing resource allocation in various production environments.