关键词: 产品族, 神经网络, 绩效控制, 配置设计, 学习效应, 规模效应

Abstract: To deal with the issue of performance optimization and control in the gradual evolution process of modular product family for make-to-order enterprises,the neural network and configuration optimization techniques were utilized to realize performance maximization.Both abrupt change and gradual evolution in product family evolution process were elaborated,meanwhile certainty and uncertainty of dynamic market requirement in gradual evolution process were analyzed.Performance optimization and control framework for dynamic implementation phase of product family was established,and the influence of product configuration design for random orders on product family cost,revenue and total performance was analyzed by comprehensive consideration of learning effect and scale effect from component commonality.The configuration design model was put forward with the goal of performance optimization and the constraints of customized requirements,product structure and cost requirement,and was solved by genetic algorithm.A case study of electric scissors product family was carried out to verify the effectiveness of proposed method,and the sensitivity of critical variables or parameters in the model was analyzed.The conclusion showed that the maximum performance in certain period of product family dynamic implementation process was realized with random orders oriented configuration design optimization,and the configuration scheme of random orders with consideration of performance factors should be confirmed.

Key words: product family, neural network, performance control, configuration design, learning effect, scale effect