Maximizing Throughout: Leveraging Meta-heuristics for Plant Scheduling

As a part of the Analytics and Insights team at Tata Consultancy Services (TCS), Anuj specializes in supply chain management, operation management, and optimization. He has worked on various supply chain and manufacturing domain projects. He has a PhD in industrial engineering from Indian Institute of Technology (IIT) Delhi. Anuj has worked on applying various meta-heuristics and fuzzy systems to flexible system problems. He has published papers in various international journals, and also presented them at conferences. Avneet is a subject matter expert in TCS' Analytics and Insights practice. He holds a PhD in supply chain and simulation from IIT Delhi. He has more than 14 years of industry, research, training, and consulting experience. His experience includes projects in complex simulation, optimization, and statistical and business analytics. He is a certified PMI-PMP, six sigma, and APICS supply chain professional. He has published several papers in leading research journals. Production scheduling, or allocation of machines and resources to operations, plays a critical role in planning and implementing an efficient, flexible, and profitable manufacturing system. It is an operational level decision, which helps optimize capacity utilization and maximize output when executed systematically. The complexity of scheduling in a plant increases due to machine flexibilities and precedence relationships between operations within a job. Hence, traditional techniques such as Johnson's n/2 method, priority rules, and other heuristics might be insufficient for providing an optimized schedule. In such a complex situation, meta-heuristic optimization is a dynamic technique for effective and efficient decision making. This paper discusses the various issues associated with scheduling in a manufacturing environment, and how best to resolve them. Our proposed solution is a scheduler built on a set of meta-heuristics (MH) for optimal scheduling. We have used a numerical example to illustrate the efficacy of the proposed algorithm. On comparing the results with those obtained from applying traditional priority rules, we found that the proposed approach achieves greater optimization. Abstract Contents