Disruption Management For Permutation Flowshop Scheduling Considering Real-world Behavioral Participators

Permutation Flowshop Scheduling Problem (PFSP) has profound practical background and a wide variety of applications in practice, and PFSP is also a classical combinatorial optimization problem. In generally, PFSP is devoted to solve production scheduling problem in an ideal manufacturing environment. That is to say, there is no disruptions, and it is assumed that the behavioral participators in the production processing system is "completely rational". However, the scheduling systems usually have effected by many disruption events in practice, such as machine breakdown, new arrival orders, changes of jobs’priorities, et al. Disruption management focus on how to recover from the negative effects of disruptions at costs as low as possible, as a management science frontier.New arrival orders is one of the most common disruption events in the manufacturing system. For that, the research focuses on the predictive-reactive scheduling disruption management in the permutation flowshop environment. In addition, behavioral decision science research shows that the decision-maker is often in a state of "limited rationality", and the risk of unknown, the uncertainty of information, the complexity of things, the initial goal to change can limit the subject to make rational decision, which lead to serious disconnection between theoretical research and practical applications. Given that the classical production operation management research is based on the hypothesis of man’s "complete rationality", which is unreasonable in practice, the research explores new disruption measures around the disruption management for the PFSP from the perspective of the behavioral participators.Finally, disruption management model is built up, considering the original objective and disruption objective. As for the advantages and disadvantages of existing algorithms, taking the existing algorithm improvement methods into account, we design the problem solving algorithms and do numerical test.The main endeavors and contributions of the dissertation are as follows:(1) In permutation flowshop scheduling environment, the disruption effect of new arrival orders is analyzed from the perspective of the main behavioral participators including enterprise manager, shop worker and customer. The unsatisfactory degree function based on prospect theory is designed to quantify the effect of the disruption. Further the bi-objective disruption management model is built up. (2) For the bi-objective disruption management model, a general hybrid strategy based on the advantages and disadvantages of the existing intelligent optimization algorithms, is proposed for improving initial solutions and strengthening local search. Then typical meta-heuristic algorithms are chosen from algorithms based on population and trajectory methods based on one single solution for hybrid algorithms to design.(3) We design computational study by integrating benchmark flowshop testing dataset Taillard (Ta) and random generation of experimental data. The advantage of proposed deviation measure based on prospect theory over existing measures is shown. By analysis of proximity and diversity of obtained Pareto front, the proposed hybridization strategy is as well demonstrated to be effective.Finally, to summarize the research, and point out that the limits and future research directions.