大规模定制下复杂产品配置设计交互协同优化研究

As a multi-level coupling body with highly integrated structure and function, the core technology of complex product is fast related to national development and people's livelihood. Under the scale of mass customization, there are problems exist in the field of complex product configuration, like the representation methods for product configuration knowledge, or solving strategy of the configuration model. To solve these problems, this study focuses on collaborative optimization of complex product configuration design in mass customization. Detailed researches are carried out on the following issues: representation strategy for complex product configuration, product configuration algorithm, coupling optimization of complex product configuration and conceptual model of the complex product configurators. In order to improve the capacity of integrated automation and intelligence of China's mass customization oriented complex modern product design, this study strive to provide the basic research theories and methods of universal significance, and to provide accurate and reliable basis for the development of complex product systems.. Main Research Contents: 1.Configuration knowledge representation model based on the CBR multi-level tree.2.Improve the global sensitivity equations (GSE), establish redundant coupling abatement strategies for complex products.3.Solving method for complex product configuration based using Interactive Genetic Algorithm (IGA) based on large population size.4.Content integration & complex product configuration system architecture design under mass customization.. Innovation of this project:1.Skeleton instance model is introduced to capture and define the design intent and product mix. A complex product collaborative configuration development oriented multi-level tree mapping system for knowledge representation is designed. This study aims to create a case-based reasoning-centered design process model based on instance and complex structural constraints.2.An improved collaborative interactive genetic algorithm based on non-fixed-length tree coding is proposed to solve complex product configuration problems and en efficient solving process model is established for complex product configuration design, with the integration of neural network generalization ability and the use of non-user-specified individual fitness, large-scale population interactive genetic algorithm.3.Content of the project will be integrated to better our plan ability on complex product configuration under mass customization. Complex product configuration system architecture is designed for Mass Customization. Implementation strategies and schemes based on the facts of manufacturing enterprises in China are proposed.

复杂产品系统是一个结构和功能高度集成的多层次耦合体。针对大规模定制下复杂产品配置的知识表示和配置问题求解等研究存在的问题，重点研究以有效的产品表示方法为基础，考虑复杂产品配置耦合重叠问题，设计适用于大规模定制的复杂产品配置优化方法。内容包括：基于实例推理，引入骨架实例模型，捕捉并定义设计意图和产品结构，同时分析其子系统间耦合关系，修正全局灵敏度方程，拟提出复杂产品配置冗余耦合变量消减策略，研究面向复杂产品协同配置开发的多级树知识表示模型；研究大规模种群交互式遗传算法并融合神经网络泛化功能，提出一种改进的基于非定长树编码的协同交互式遗传算法解决复杂产品配置问题；整合项目研究成果，规划大规模定制下复杂产品配置系统。同时，对汽车设计制造企业的大规模定制问题进行建模，通过仿真，验证本项目提出的理论与方法体系的先进性和合理性。研究成果将为大规模定制下复杂产品系统创新开发提供理论科学依据。

研究主要针对大规模定制背景下复杂产品协同定制配置设计问题，重点研究以有效的产品表示方法为基础，考虑复杂产品配置耦合重叠问题，设计适用于大规模定制的复杂产品配置优化方法。包括产品配置模型、面向客户需求的产品配置方式、产品配置知识表示、复杂产品配置空间分配、兼容性约束和选择性约束处理、产品耦合情况处理、交互式遗传算法及用户疲劳和犹豫感分析，汽车操控台概念设计等关键问题和技术，多学科交叉结合，运用管理科学、机械设计学、人工智能学、计算机科学、运筹学等学科的理论和方法，通过建模、求解、仿真等过程给出产品的设计方案，以满足客户对产品日益膨胀的个性化需求，并对所提出的模型和改进的算法进行验证。本研究对于复杂产品的协同配置概念设计具有重要的理论意义，对于制造业企业快速、准确获取客户需求，进行新产品的研发和设计具有重要的实践意义。具体研究成果包括：提出改进的交互式遗传算法，进行产品配置设计，采取不同形式的知识表示模型，以汽车操控台为例，进行新产品的研发；进行PIC模型下大规模定制的产品配置方法，建立基于PIC模型的汽车操控台定制设计系统，验证PIC模型对于产品知识的表示具有良好的效果；将人工神经网络和交互式遗传算法相结合，以另一种复杂产品——智能手机为实例，验证该算法在产品配置中的性能，以测试算法的延展性；进行云环境下无线多媒体社交网络的带宽分配进行相关研究，结合基于Qos的Web服务组合方法，为云环境下产品的协同配置设计研究提供重要的理论研究基础。涵盖了申请书中所有研究内容。共发表科研论文13篇，其中SCI收录8篇次，国内核心期刊论文3篇，EI检索和会议论文2篇。项目在研期间，申请人兼任多个期刊编辑。此外，申请人还积极参与学术会议活动，分别担任第44届计算机与工业工程国际会议分会主席，第45届计算机与工业工程国际会议委员会委员兼分会主席。