Research On Optimal Scheduling Of Hybrid Production Processes

Generally we can divide production processes into three categories, they are discrete, continuous and hybrid. This dissertation focuses on production scheduling methods of hybrid production processes.In the dissertation, we first state general production planning and scheduling problems and drawbacks of classical solutions. Then, a theoretical decomposition-based approach is introduced to solve the main issue of inefficient capacity utilization. To optimize multi-stages' hybrid production processes, we first study the optimal scheduling of discrete and continuous production processes, and then we use the results to further study the optimal scheduling of hybrid production processes.In discrete production processes, we studied production scheduling problems of two identical parallel machines at single stage and multistage. The objective is to minimize production bottlenecks and the penalties of Earliness/Tardiness (E/T). We used heuristics branch and bound algorithm in this study since scheduling of parallel machines is a NP-hard problem and we proposed four heuristics optimal scheduling algorithms in this study. We also did simulation using our proposed methodology to prove the effectiveness of the algorithms.In studying continuous production processes, we considered the problem of batch processing scheduling where each processor can process a batch of jobs simultaneously; the batch is delivered to the customer together upon the completionof the batch. In JIT scheduling, the objective is to minimize the penalties of E/T, and the cost of processing. In order to determine an optimal batch starting and dispatching, we proposed a heuristics optimal scheduling algorithm and we showed the effectiveness of our algorithm numerically.Inventory management is very important in multistage hybrid production processes. According to the demand of the stochastic model of inventory for the single stage, we analyzed modeling methods of multi-stage production/inventory system in hybrid production processes. Then we introduced a model of inventory system for multi-stage processes. Finally, we introduced an evaluation method for the parameter of the models.Based on our analysis of discrete and continuous production processes, we study the optimal scheduling method in the hybrid production systems with the objective to minimize the sum of production costs. To overcome a complex scheduling problem, we introduced a model that divided scheduling into two layers. At the upper layer of the model, we focus on batch job scheduling, precisely when and where jobs will be produced in hybrid production systems, while at the lower layer, the elements of the batch are scheduled and built in a hybrid production system. Based on our extensive study of many different scheduling optimizations, we can summarize the results into five scheduling rules. Based on the rules and the heuristics branch and bound algorithm we can obtain an optimal scheduling algorithm for both layers of our model. Finally, we applied our optimal scheduling algorithm in the Zircotube productive system of Northwest Zircotube Co. Ltd., which has a hybrid production system, and we achieved satisfactory results.