Lookahead Batching-based Dynamic Batch Scheduling In A Two-stage Hybrid Flow Shop

Batch scheduling problem is a challenge for nearly all aspects of industry,such as chip pre-burning processing in semiconductor manufacturing and heat treatment in metal processing and etc.Compared with the classical scheduling problem,batch scheduling is much more complex because there is no exclusive assumption of the processing.The flow-shop environment is more practical shop configuration,so it is significant both in theoretical research and practical implementation.The batch scheduling problem in the mould thermoechanical treatment workshop is a typical dynamic bach scheduling problem with a two-stage hybrid shop configuration.The characteristics of the workshop are described as follows: Each stage contains multiple batch processing machines.Non-identical jobs arrive to the shop stochastically and they are processed in the two stages in batches.Each batch needs to be disassembled and inspected for quality after each stage is completed,and multiplex rework randomly occurs,and background industry has a combined demand for tardiness,machine effectiveness,and batch quality.For the above characteristics,the main research work of this thesis is as follows:(1)A lookahead batching heuristic for uncertain environments is proposed,which can enable batch optimization of the two stages and fast response to rework interference.In the heuristic,an event-driven lookahead bacthing decision can determine the next best batch scenario for the currently idle machine in the first stage.The bacth scheduling for the second stage is triggered by linkage at the same time,and the original scheduling is properly adjusted once the rework interference appears.The core of the developed heuristic is a two-level compromise programming,which is capable of deep coordination of the bi-objective of batch tardiness index and batch quality index.Simulation results show the presented heuristic outperforms five similar methods.(2)Considering the limitations and short-sightedness of the lookahead batch decision in global scheduling,the pre-process pulling mechanism is embedded into the lookahead batching heuristic,a new lookahead batching heuristic is proposed.The quenching batch decision determines the processing sequence of the roughing process and improves the batching optimization results.Simulation results show the presented heuristic has better performance than the heuristic without the pre-process scheduling decision mechanism.(3)Environmental adaptability simulation experiments are conducted in comparison with two heuristic proposed in this thesis and other similar heuristic.By changing four environmental factors,we research on the adaptability and limitations of the two heuristic compared with other algorithms in different environments.