【作者】 张利平；

【导师】 高亮； 李培根；

【作者基本信息】 华中科技大学 ， 工业工程， 2013， 博士

【摘要】 随着全球市场竞争的加剧、客户需求的多样化与个性化，实际制造环境中紧急件、工件随机到达等状态频发，使得动态调度问题成为制造系统研究领域的热点之一。同时，随着经济的快速发展，环境问题日益被关注。因此，本文围绕动态环境下作业车间、柔性作业车间调度问题以及面向低碳运行的多目标柔性作业车间调度问题展开研究。实际的制造车间中，往往存在很多动态不确定因素，如何保持车间生产稳定以及车间生产效率，制定合理的调度计划是非常重要的。本文结合传统作业车间调度问题的整数规划模型、动态调度问题特性与动态事件属性，建立了作业车间动态调度的数学模型。为了提高预-反应调度的鲁棒性，设计了一种带空闲时间的预-反应调度策略。实验结果表明，对于小规模生产车间，带空闲时间的预-反应调度策略可保证原始调度的鲁棒性且增强重调度的效率和稳定性；但对于大规模生产车间，完全重调度策略具有较好的调度性能。实际制造系统中，有效的重调度求解方法对决策者进行决策具有重要影响。在充分考虑车间生产信息的基础上，本文结合遗传算法的全局搜索能力以及禁忌搜索算法的局部搜索能力，提出一种具有新初始化方法的遗传禁忌搜索算法求解单目标作业车间动态调度问题。实验结果表明，新初始化方法能保持种群多样性和提高算法的全局搜索能力，且提出的遗传禁忌搜索算法具有良好的鲁棒性。在实际车间环境中，如何权衡调度效率和调度稳定性两方面的问题是解决动态调度问题的关键。本文建立了基于调度效率和调度稳定性的多目标作业车间动态调度模型，设计了基于遗传禁忌搜索算法的动态调度求解方法，在该方法中，任何一个调度周期内，发生器为下一阶段产生动态事件，遗传禁忌搜索算法优化问题并产生预调度方案。实例验证了提出的模型和求解方法的有效性与优越性。针对柔性作业车间动态调度系统，本文在传统指标中添加了平均工序数量参数，完善了柔性作业车间动态调度的评价指标体系，建立了柔性作业车间动态调度模型，设计了一种有效的遗传变邻域算法，实例验证了该方法的有效性。采用实验手段探讨了调度周期对调度系统的影响，结果表明不同车间负荷水平下的调度周期与调度系统性能均大致呈U型曲线。采用ANOVA方法进行统计显著性试验，结果表明车间负荷水平、新工件数量均对调度效率和调度稳定性具有统计显著影响。随着全球气候变暖与经济快速增长，降低能耗成为国际政治、经济与学术研究关注的热点之一。首先，利用机床加工工序过程的总空载能耗以及能量利用率等公式，设计了用于动态调度的基于工序加工时空载能耗的评价指标。接着，采用两种方式研究了面向低碳运行的多目标柔性作业车间动态调度问题。一是，建立了面向低碳运行的能耗与调度效率的目标规划模型，设计了带精英策略的遗传算法的求解方法，实验结果表明，最小能耗模型可有效降低能耗，且在一定程度上提高调度效率。二是，建立了综合考虑能耗、调度效率与调度稳定性的多目标柔性作业车间动态调度模型，实验结果表明该模型可有效降低能耗，保证调度效率和调度稳定，这对应对气候变暖、要求低碳运行的大环境是有益的。最后，采用ANOVA方法进行统计显著性试验，结果表明车间负荷水平、新工件数量均对能耗、调度效率与调度稳定性具有统计显著影响。在以上研究成果的基础上，根据某发动机冷却风扇加工车间的实际情况，分析了该车间存在的问题，将上述理论成果应用于实际车间生产进行了实例测试与分析。最后，对全文工作进行总结，并对今后研究方向进行展望。更多还原

【Abstract】 With the increasing competition of global market and the diversity of customerdemand, many dynamic events have been occured in the real manufacturing. The dynamicscheduling problems have become one of the hot topics in the field of manufacturingsystem. Moreover, with the increasing rapid development of global economy,environmental problems are becoming the focus of attention. As a new sustainablemanufacturing mode, low carbon manufacturing is an effective way to realize the pledgefor conserving energy and reducing emissions in our country by2012. This paper focuseson the research of the dynamic job shop scheduling problem, the dynamic flexible jobshop scheduling problem and the multi-objective dynamic flexible job shop schedulingproblem based on low carbon.Dynamic events always occur in the real shop floor. It is important to keep thescheduling stability, improve the scheduling efficiency and make reasonable schedulingdecisions. According with the integer programming for the classical job shop schedulingproblem and the characteristic of the dynamic events, a mathematical model for thedynamic job shop scheduling problem is proposed. In order to improve the robust of thepredictive/active scheduling strategy, a predictive/active scheduling strategy with insertingidle time is proposed. The experimental results show that the proposed strategy can keepthe robust of the original schedule and improve the scheduling efficiency and thescheduling stability in small scale shop floor. However, the complete rescheduling strategyhas a better scheduling performance in large scale shop floor.An effective rescheduling approach has important impact on making decisions in realmanufacturing system. One hybrid algorithm which is mixed by the genetic algorithmwith strong global searching ability and tabu search with strong local searching ability hasbeen proposed to solving the dynamic job shop scheduling problem. The experimentalresults show that the new initialization can keep the population diversity and improve theglobal search ability. They also show that the hybrid genetic algorithm and tabu searchalgorithm has the good robustness.How to balance the scheduling efficiency and the scheduling stability is a keyproblem to solving the dynamic job shop problem in real shop floor. In this paper, amulti-objective mathematical model for dynamic job shop scheduling problem, whichcontains the scheduling efficiency and the scheduling stability, has been proposed. Ahybrid genetic algorithm and tabu search algorithm is proposed to solve themulti-objective dynamic job shop scheduling problem. The simulator generates the dynamic events for next phase at each rescheduling point. The hybrid algorithm optimizesthe problem and generates the prediction schedules. The experimental results show theeffectiveness and advantage of the proposed model and the proposed approach.Dynamic flexible job shop scheduling problem is one of the extension of dynamic jobshop scheduling problem. This paper inserts a mean quantity of operations to improve theevaluation system of dynamic flexible job shop scheduling problem. The mathematicalmodel for dynamic flexible job shop scheduling problem has been proposed. An effectivealgorithm, which mixes the genetic algorithm and variable neighborhood search algorithm,has been proposed to solve the multi-objective dynamic flexible job shop schedulingproblem. The experimental results show the feasible of the proposed approach. Theexperimental results also reveal that the curve of the shop load level and the schedulingperformance is U-shape. The ANOVA method shows that the shop load level, new jobarrivals has a statistical influence on the scheduling efficiency and the scheduling stability.With global warming and the increasing rapid development of global economy,reducing energy consumption has become one of the hot topics in the research ofinternational politics, economy and the academic research. This paper designs the energyconsumption according to the operation-based processing unload energy consumption. Agoal programming model based on low carbon has been proposed. An improved geneticalgorithm with elitist strategy has been proposed to solve dynamic flexible job shopscheduling problem. The experimental results show that the minimization energyconsumption model can reduce the energy consumption and enhance the schedulingefficiency partly. Moreover, a multi-objective dynamic flexible job shop scheduling model,which contains the energy consumption, the scheduling efficiency and the schedulingstability, has been proposed. The experimental results show that the proposed model canreduce the energy consumption, improve the scheduling efficiency and keep thescheduling stability. Finally, the ANOVA method shows that the shop load level, new jobarrivals has a statistical influence on the scheduling efficiency and the scheduling stability.Based on the research word mentioned above and the real conditions in the enginecooling fan shop floor, the issues in the shop floor have been analyzed. The above researchresults have been applied into the shop floor to test and analyze.Finally, the research results achieved in the dissertation is summarized and futurework is generalized and looked forward.更多还原