【作者】 曹鹏彬；

【导师】 肖人彬；

【作者基本信息】 华中科技大学 ， 机械设计及理论， 2007， 博士

【摘要】 产品设计方法在产品开发中起着重要作用。生命现象蕴涵着无穷的奇妙和灵巧,可为包括产品设计在内的各种复杂问题的有效解决提供启迪和灵感。近年来,随着各种仿生计算技术的不断兴起和发展,基于各种仿生计算方法的产品设计方法不断涌现并取得了令人瞩目的成功。人工免疫系统(Artificial Immune System, AIS)是受生物免疫系统启发而产生的一种新兴仿生计算方法,它提供了一种强大的信息处理和问题求解范式,可为产品设计提供新的智能支持手段。为此,本文研究了基于人工免疫系统的产品设计方法,旨在探索产品设计的新思路和新方法。面向产品设计的三维空间,即产品设计层次(包括设计问题求解和产品演化机制两个层次)维、设计进程维和产品设计方法维,建立了基于AIS的产品设计框架,旨在为基于AIS进行产品设计提供框架指导。此外,基于AIS解决工程中各种实际问题所具有的共性特点和聚类特征,概括提出了工程免疫计算(Engineering Immune Computing, EIC)的概念,进而在设计问题求解层次构建了基于EIC的产品设计框架作为基于AIS的产品设计框架的一个子框架,以此指导产品设计中的设计问题求解。鉴于基于AIS的产品设计涉及的范围广泛,无法一一展开研究,因而本文在基于AIS的产品设计框架指导下,从一些有代表性的角度出发,开展了相关研究工作。首先,研究了基于EIC的产品功能规划方法。该方法为了克服现有公理设计在处理耦合设计方面存在的不足,通过功能辨识(基于分解操作或免疫聚类学习方法)、功能耦合程度度量(采用两两比较方法并基于三角模糊数进行度量)以及耦合功能规划(包括解耦和割裂两个操作,割裂实际上可以归结为优化问题,可采用枚举割裂或免疫割裂方法实现),为公理设计过程中出现的耦合设计问题提供功能规划的解决方案。此外,以冷凝器的功能规划为例,对所提方法进行了说明和验证。该部分研究工作探讨了EIC在产品设计中出现的聚类学习和优化两类典型设计问题求解中的应用,说明了EIC可为产品的上游概念设计阶段以及公理设计方法提供有力支持。其次,研究提出了一种面向装配规划问题求解的新方法——基于免疫优化的装配规划方法,不仅能够对产品的装配序列进行优化,还能够对装配操作的工艺进行优化,从而实现了产品装配过程的优化。通过引入多种免疫操作,基于免疫优化的装配规划方法能够较好地实现兼顾全局搜索和局部搜索的均衡搜索,还能够有效处理约束和利用装配的启发式经验知识为装配规划提供指导,从而获得较好的求解质量和较高的装配规划效率。此外,通过比较详细的实例研究,说明了所提方法对于求解装配规划问题的有效性和优越性。该部分研究工作探讨了EIC在产品设计中广泛存在的优化问题这类典型设计问题求解中的应用,同时说明了EIC可为产品的下游工艺设计阶段以及DFA方法提供有力支持。然后,研究了基于AIS的创新设计原理。从信息处理和计算的角度,对AIS所具有的优良特性进行了阐述,进而指出创新设计可以采用创新计算的途径来研究实现,由此分析提出了基于AIS的创新设计构想。此外,建立了一个以免疫创新设计方法(包括面向问题求解的免疫辅助创新设计方法和面向产品演化的免疫设计方法)为核心的基于AIS的创新设计框架,并对主要关键技术进行了讨论说明。该部分研究工作一方面借鉴生物免疫系统运行机制中所蕴涵的计算模式和创新模式,为创新设计提供了一条新颖途径;另一方面说明了AIS可从设计问题求解和产品演化机制两个不同层次为产品创新设计方法提供有力支持。此外,基于产品功能规划的研究成果,开发了基于EIC的产品功能规划原型系统,验证了所提方法的正确性和有效性。更多还原

【Abstract】 Product design method plays an essential role in product development. Since the phenomenon of life implies infinite strangeness and dexterity, it can provide eligntenments and inspirations in efficient settlement of various complicated problems including product design. In recent years, as the rise and progress of diversified bionic computing techniques, product design methods based on various bionic computing techniques have come forth and acquired noticeable success. As a kind of emerging bionic computing technique that draws inspiration from the vertebrate immune system, Artificial Immune System (AIS) offers a powerful information processing and problem solving paradigm, hence it is natural to view AIS as a new means that provides intelligent supports for poduct design. In this paper, the product design method based on AIS is examined with the purpose to explore novel ideas and methods for product design.Aiming at providing an overall direction for carrying out product design based on AIS, a framework of product design based on AIS is established against the three-dimensional product design space, which consists of the dimension of hierarchies of product design composed of the hierarchy of design problem solving and that of product evolutive mechanism, the dimension of design process and the dimension of product design methods. In addition, according to the analysis of the common and clustering features of using AIS to solve engineering practical problems, the concept of engineering immune computing (EIC) is abstracted and generalized, and at the hierarchy of design problem solving, a framework of product design based on EIC is established as a sub-framework of the framework of product design based on AIS to direct design problems solving in product design.Because product design based on AIS covers a broad scope, it is impossible to carry out investigation for its each aspect. Thus in this paper, some representative aspects from the framework of product design based on AIS are selected as breakthrough points to be studied in detail.Firstly, a method of product function planning based on EIC is investigated to overcome the limitation that existing Axiomatic Design theory cannot deal with coupled design. The aim of the method is to offer function planning solutions for coupled design problems that appear in the procedure in design with Axiomatic Design, which is actualized by function identification based on the partition operation or immune clustering learning, measurement of the coupling strengths of coupled functions using the pair wise comparison technique based on triangular fuzzy number, and coupled function planning including the decoupling and the tearing operations. In reality the tearing operation can be considered as an optimization problem and can be achieved by the enumeration method or immune tearing method. In addition, the function planning of a condenser is taken as an example to illuminate and validate the proposed method. This part of study of the paper explores the application of EIC in solving two typical design problems, namely clustering learning and optimization, and shows that EIC can provide convincing supports for the conceptual design in the upriver phase of product design and the product design method of Axiomatic Design.Secondly, a novel approach called immune optimization approach (IOA) is studied and presented to solve the assembly planning problem. IOA not only can optimize the assembly sequences of products, but also can optimize the process of assembly operations. Therefore IOA optimizes the assembly procedures of products in reality. By introduction of manifold immune operations, IOA achieves both global and local searches, i.e. balanced search is achieved. Besides IOA also can deal with constraints efficiently, and can use heuristic experiences and knowledge on assembly to give guidance to assembly planning. For the above reasons, IOA can obtain good assembly sequences with satisfying efficiency. To demonstrate the validity and advantages of assembly planning using IOA, detailed case study is presented. This part of study of the paper explores the application of EIC in solving optimization problem, a widely existing design problem in product design, and shows that EIC can provide convincing supports for the process design in the downriver phase of product design and the product design method of DFA.Thirdly, the principle of creative design based on AIS is investigated. From the point of view of information processing and computation, the main excellent characteristics held by AIS are expounded. Then it is pointed out that creative design can be achieved by the avenue of creative computation, and the idea of conducting creative design based on AIS is analyzed and presented. Furthermore a framework of creative design based on AIS is established with immune creative design methods, involving the immunity aided creative design method for problem solving and the immune design method for product evolution serve as the kernel. The key techniques of implementing creative design based on AIS are also discussed. This part of study of the paper on the one hand offers a novel avenue to creative design by borrowing the computational and the creative patterns from the operational mechanism of the vertebrate immune system. One the other hand it shows that AIS can provide convincing supports for the product design method of creative design both from the design problem solving hierarchy and the product evolutive mechanism one.Finally, based on the research fruit of product function planning, a prototype system of product function planning based on EIC is developed to validate the correctness and the validity of the proposed method.更多还原