【摘要】 探讨板式家具生产在缓冲约束下的混合流水车间调度问题，建立缓冲约束，并研究求解方法，为解决由于当前家具生产调度方法缺乏考虑缓冲约束使得现代调度技术难以实际应用的问题提供科学依据。以板件数量作为缓冲约束中容量的表征，根据混合流水车间调度问题的特征，建立工序间有限缓冲约束，并将其编码进遗传算法的适应度函数中；设计满足调度问题特征的交叉操作、变异操作、个体评估与选择操作。其中，交叉操作采用部分映射法，变异操作采用单点插入法，个体评估采用已建立的适应度函数，选择操作则采用精英保留策略和轮盘赌方法。最后利用MATLAB对遗传算法各模块进行编程，通过文献中的案例进行算法的可行性验证。通过对已有文献的调度规则和方法(先进先出原则、NEH算法、模拟退火算法、粒子群算法、蚁群优化算法和改进布谷鸟搜索算法)进行对比试验，结果显示本研究提出的遗传算法在以完工时间为优化目标的前提下均优于其他方法。同样在考虑缓冲约束的案例场景中，本研究提出的方法也具有有效性。基于遗传算法的工序间有限缓冲约束下板式家具多产线混合流水车间调度问题的结果具有一定的可行性，可以为板式家具生产调度技术提供新的解决思路，但仍需综合考虑更多的生产动态因素来提高其实际应用能力。更多还原

【Abstract】 Job shop scheduling technologies have significantly improved the production processes of panel furniture by shortening the production cycles, leading to enlarged profit margin of enterprises. The existing job shop approaches for panel furniture production assume unlimited storage spaces between production processes, but this is not practically the case and thus limiting the efficiency of these approaches for real-world applications. This research discussed the hybrid flow shop scheduling problems with limited buffer capacity in the panel furniture production, established the constraint of limited buffer capacity, and investigated the solution methods, which provided a scientific basis for solving the problems that modern scheduling technologies were difficult to be applied in practical manufactures due to the lack of consideration of buffer capacity constraints in current furniture production scheduling methods. Taking the number of plates as the representation of cache capacity, according to the characteristics of the hybrid flow shop scheduling problem, the constraint based on the limited cache capacity between processes was established, and it was encoded into the fitness function of the genetic algorithm, and then the crossover operation, variation operation and individual evaluation and selection operation were designed to meet the characteristics of scheduling problems. Among them, the crossover operation adopted the partial mapping method, the mutation operation adopted the single point insertion method, the individual evaluation adopted the previously established fitness function, and the selection operation adopted the elite retention strategy and roulette method. Finally, MATLAB was used to program each module of the genetic algorithm, and then the feasibility of the algorithm was verified by the reported cases in the literature. The comparative tests of some scheduling rules and methods in the existing literature, such as the first in first out principle, NEH algorithm, simulated annealing algorithm, particle swarm optimization algorithm, ant colony optimization algorithm and improved cuckoo search algorithm were carried out. The results showed that the genetic algorithm proposed in this study was due to other methods on the premise that the completion time was the optimization target. Notably, it was shown about a 27.9% decrease in production cost compared with the first in first out rule that is commonly used in the furniture manufacturing industry. In the case scenario considering cache constraints, the scheduling method proposed in this research was effective. The result of the hybrid flow shop scheduling problem of panel furniture with limited buffer capacity between processes based on genetic algorithm had certain feasibility and could provide a new solution for the production scheduling technology of panel furniture, improving its practical application ability. However, it is expected to comprehensively consider more production dynamic factors in the future research.更多还原