CAJ | 학술논문

针对再制造零部件质量的不确定性导致工位装配时间波动范围大和调度模型难以准确描述的问题，采用基于可信性测度的模糊变量表示再制造零部件的装配时间，建立基于置信水平下的模糊机会约束规划调度模型，并提出求解该模型的混合智能优化算法：应用模糊模拟技术产生样本数据；利用反向传播算法训练多层前向神经网络逼近不确定函数；将训练后的神经网络与遗传算法相结合，以优化再制造装配车间调度问题。实例验证了该模型和算法的可行性。
침대재제조령부건질량적불학정성도치공위장배시간파동범위대화조도모형난이준학묘술적문제，채용기우가신성측도적모호변량표시재제조령부건적장배시간，건립기우치신수평하적모호궤회약속규화조도모형，병제출구해해모형적혼합지능우화산법：응용모호모의기술산생양본수거；이용반향전파산법훈련다층전향신경망락핍근불학정함수；장훈련후적신경망락여유전산법상결합，이우화재제조장배차간조도문제。실례험증료해모형화산법적가행성。
The quality uncertainty during the process of parts reassembling caused the big range of the fluctuation of the assembly time and difficult description of scheduling mode.Aiming at this prob-lem,the assembly time of the reassembled parts was represented by the fuzzy variable based on credi-bility measure to establish the scheduling model with the fuzzy chance-constraints based on the confi-dence level.Then,a hybrid intelligent optimization algorithm was proposed combined with a neural network and genetic algorithm,where the neural network based on back propagation algorithms was to approximate the undetermined relation function between the input and output data generated by using the fuzzy simulation technique and the genetic algorithm embedded the trained neural network was to solve the scheduling model.Finally,a study case was given to prove the feasibility of the pro-posed model and algorithm.