Intelligent Optimization Methods For Flowline Manufacturing Cell Scheduling

The cellular manufacturing system is an effective producing system which applies the group technology for flexible batch production environments. The flowline manufac-turing cell scheduling in this environment belongs to machine scheduling in which there is no resource-constraints. This kind of problems is widespread in flow type scheduling production environments, such as chemical, metal, and food processing. Optimizing such problems leads to reducing job processing time, decreasing the work-in-process inventories and improving the quality productions.Flow shop scheduling with multiple machines has been proved to be a NP-hard prob-lem. Exact methods are applicable only for problem with small to medium sizes because of the exponential time complexities. For problems with large scale, heuristic or meta-heuristic methods are commonly used to fund solutions in reasonable running times. Some classical flowline cellular manufacturing system scheduling problems are considered in this dissertation with major contributions shown below.(1) For the no-wait flowline scheduling problem with total flowtime minimization, a composite heuristic is proposed based on the insertion-segment optimization. By analyz-ing the properties of objective increment and considering the characters of current existing heuristics, an insertion-segment neighborhood searching framework is proposed. Iterative pair-exchange method is applied to enhance the quality of the solution. Experimental results show that the proposed algorithm performs steadily in effectiveness and meets the requirements in real-time production environments.(2) For the flowline manufacturing cell’scheduling problem with sequence-dependent family setup times to minimize makespan, a hybrid meta-heuristic is proposed based on harmony search algorithm. The characters of flowline cellular manufacturing system are analyzed. A two-level scheduling strategy is used. Iterative optimization algorithm, which is composed of RZ and PE optimization operators, is applied for obtaining better solutions. A simple method with discarding the worst solutions is proposed to avoid the algorithm trapping into premature and local optimization. Experimental results show that the proposed algorithm can efficiently improve the quality of the solutions, which is suitable for the production environment with low-level real-time requirements.(3) For the flowline manufacturing cell scheduling problem with sequence-dependent family setup times to minimize total tardiness and mean total flowtime, a bi-objective hybrid meta-heuristic based on harmony search algorithm is proposed. According to the nature of this problem, a two-level flowline scheduling strategy is used. An iterative optimization algorithm with RZ optimization operator is applied for obtaining better solutions. In order to avoid premature and local optimization of the algorithm, a crossover operator is presented during the variation process which enables the algorithm to skip the searched area with local optimization and search more solution space. Experimental results show that the proposed algorithm can obtain more Pareto-based solutions with better quality.(4) For the flowline manufacturing cell scheduling problem with sequence-dependent family setup times which considers inter-cell moves to obtain makespan minimization, a composite heuristic and two meta-heuristic algorithms are proposed. For the composite heuristic, a method similar to NEH algorithm is used to generate a candidate solution for the heuristic algorithm. A local optimization procedure is deployed to enhance the quality of the solution until no improvement could be achieved. For the proposed hybrid harmony search algorithm and hybrid genetic algorithm, the local search strategy is also used to obtain better quality schedules. Experimental results show that the composite heuristic is good in efficiency and the hybrid genetic algorithm obtains good solutions.