New solution approaches for job scheduling in virtual manufacturing cells

In today?s competitive market conditions, shortening product life cycle and varying demand pattern create a highly dynamic environment. Firms have to explore and adapt to novel production systems for surviving in the volatile conditions and reducing their costs. Especially offered for the companies that perform small-to-medium batch production, Virtual Manufacturing Cells (VMCs) are prominent among these relatively new manufacturing systems. VMCs appear to be a hybrid form accommodating the flexibility feature of Flexible Manufacturing Systems (FMSs) and basic flow-line process of Cellular Manufacturing Systems (CMSs). In this thesis, unlike existing literature, how jobs can be scheduled in a more efficient manner is thoroughly examined to reflect the realistic performance of VMCs. In the proposed VMCs scheduling problem, jobs are produced in batches and batch sizes can be divided into sub lots with smaller quantity of jobs to obtain lower Cmax value. There are multiple jobs with different processing routes and a set of eligible machines is available to process the operations. Machines are located to different areas in the shop floor to quickly respond to demand changes. This, however, arises the issue of travelling time consideration. A Mixed Integer Linear Programming (MILP) model is developed for the problem defined above. Due to the intractability of the problem, MILP model suffers to provide solutions for big sized problems in a reasonable amount of time. We therefore present a Genetic Algorithm (GA) heuristic approach with four vectors to obtain satisfactory results in a shorter computational time. On a wide range of randomly generated test instances consisting of 720 problems, comparative results show that GA is quite favourable and it finds the optimum solution for all problems in the case that sub lot number is equal to 1.