Improved Heuristics for Manufacturing Scheduling A Thesis Submitted In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Engineering By SAPKAL SAGAR ULHAS Under the Supervision of Dr. DIPAK LAHA Mechanical Engineering Department Jadavpur University Kolkata - 700032, India September 2012
Improved Heuristics for Manufacturing Scheduling A Synopsis Submitted In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Engineering By SAPKAL SAGAR ULHAS Under the Supervision of Dr. DIPAK LAHA Mechanical Engineering Department Jadavpur University Kolkata - 700032, India September 2012
Improved Heuristics for Manufacturing Scheduling SYNOPSIS In the manufacturing context, scheduling is concerned with setting the order in which a given number of jobs must be processed on a specified set of machines so that one or more decision objectives are optimized. Flow shop scheduling problem is a class of widely studied scheduling problem with a strong engineering background, which represents nearly a quarter of manufacturing systems, assembly lines and information service facilities nowadays. Recent developments in scheduling theory have focused on extending the flow shop models to include more practical constraints. An important kind of flow shop scheduling problem is characterized by a no-wait constraint. A flow shop scheduling problem in which each job must be processed until completion without any interruption either on or between machines is called as no-wait flow shop scheduling problem. Some typical applications of this problem are encountered in metal processing, plastic moulding, steel factories, hot rolling industries, chemical and pharmaceutical processing, and also in advanced manufacturing, such as just-in-time production. A no-wait environment arises from the characteristics of the processing technology itself or from the absence of storage capacity between operations of a job. Within these environments, the goal is to find a schedule which tries to optimize a specified objective, such as total flow time. In manufacturing environments, total flow time minimization leads to stable or uniform utilization of resources, rapid turn-around of jobs, and minimizing in-process inventory. The no-wait flow shop scheduling problem comes under the class of NP-hard problem. For solving scheduling problems, known to be NP-complete, simple exact optimization methods such as integer programming or branch and bound have the limitations of solving with only small-sized problems. As the problem size increases, the computational time of the exact methods grows exponentially. So the use of heuristic solution procedure that has a modest computational requirement becomes the necessity though it does not guarantee optimality. 1
Due to the significance of no-wait flow shop scheduling problem in theory and in engineering applications, and due to the NP-hard nature of the problem, it is important to develop effective and efficient heuristics for solving this problem. These heuristics can be categorized as constructive heuristics and metaheuristics [1]. Constructive heuristics provide fast, frequently high quality solutions and are usually much easier to be implemented than metaheuristics. A good constructive heuristic developed for scheduling problem can be extendible to more complex environments. In the present work, we proposed constructive heuristics for the problem of general no-wait flow shop scheduling with minimization of total flow time. We developed heuristic 1 in two phases [2]. The first phase is based on the conjecture that the priority of a job in the initial sequence is given by the sum of its processing times on the bottleneck machines [3]. The initial sequence is improved in the second phase through the use of job-pair called as block insertion followed by single job insertion of the block [4]. The heuristic 1 outperforms the existing two-phase heuristics [5, 6] for the problem under consideration. We proposed heuristic 2 which is developed in two phases [7]. The first phase is a penalty-shift method derived from the well-known Vogel s approximation method for generating the initial sequence [8]. The initial sequence is improved in the second phase by using the block insertion approach. The performance of heuristic 2 is superior to the existing two-phase heuristics [5, 6] and the modified existing heuristic [9]. We also developed heuristic 3 in three phases by modifying the heuristic 1. A local search technique known as forward pair-wise exchange is applied at the third phase and is looped with the second phase of the heuristic. The heuristic 3 outperforms the existing three-phase heuristics [9, 10] and is the best among all the state-of-the-art constructive heuristics for the problem under consideration. The exhaustive computational experimentations show that the proposed heuristics perform significantly better compared to the competing existing heuristics while not affecting the time complexity. Statistical tests are also used to substantiate the significance of the improved results by the proposed heuristics. On the basis of 2
computational results and statistical tests, it can be concluded that the proposed methods are relatively more effective in minimizing total flow time in no-wait flow shops. References 1. Sapkal, S.U., Laha, D., Behera, D.K., (2012) Optimization techniques for no-wait manufacturing scheduling: a review. Advanced Materials Research, Vols. 488-489, 1114-1118. 2. Sapkal, S.U., Laha, D., (2011) An improved scheduling heuristic algorithm for nowait flow shops on total flow time criterion. Proceedings of 3 rd International Conference on Electronics Computer Technology, Vol. 3, 159-163, IEEE 3. Laha, D., Sapkal, S.U., (2011) An efficient heuristic algorithm for m-machine no-wait flow shops. Lecture notes in Engineering and Computer Science, Vol. 2188, 132-136. 4. Laha, D., Chakraborty, U.K., (2009) A constructive heuristic for minimizing makespan in no-wait flow shop scheduling. International Journal of Advanced Manufacturing Technology, Vol. 41, 97 109. 5. Rajendran, C., Chaudhuri, D., (1990) Heuristic algorithms for continuous flow-shop problem. Naval Research Logistic Quarterly, Vol. 37, 695 705. 6. Bertolissi, E., (2000) Heuristic algorithm for scheduling in the no-wait flow-shop. Journal of Material Processing Technology, Vol. 107, 459 465. 7. Sapkal, S.U., Laha, D., (2012) Application of VAM to manufacturing scheduling problems. Advanced Materials Research, Vols. 488-489, 578-582. 8. Sapkal, S.U., Laha, D., (2011) Comparison of initial solutions of heuristics for nowait flow shop scheduling. Communications in Computer and Information Science, Vol. 250, 294-298, Springer-Verlag. 9. Aldowiasan, T., Allahverdi, A., (2004) New heuristics for m-machine no-wait flowshop to minimize total completion time. OMEGA International Journal of Management Science, Vol. 32, 345 352. 10. Framinan, J.M., Nagano, M.S., Moccellin, J.V., (2010) An efficient heuristic for total flowtime minimization in no-wait flowshops. International Journal of Advanced Manufacturing Technology, Vol. 46, 1049 1057. 3