中文摘要：

针对具有模糊性、缺乏系统性和主题性的新产品开发模糊前端客户创意，提出一种基于模糊物元和改进微粒群算法的混合启发式挖掘方法。首先将模糊理论引入物元分析，将客户的个性化要求、特征及相应的模糊量值结合起来建立其形式化模糊物元模型，应用模糊物元优化方法将客户多需求优化问题转换为单需求优化问题；然后给出了最优客户创意的自适应变异微粒群（AMPSO）算法的求解方法，并与遗传算法加以比较，证明该算法的有效性和先进性。最后将该算法应用于某型号汽车外观造型设计的客户创意挖掘中，有效指导了产品创新的实施。

英文摘要：

In view of the fuzzy and uncertain customer ideas in the fuzzy front end of new product development, this paper presented a hybrid heuristic algorithm based on improved PSO combined with fuzzy matter element analysis. Firstly, it integrated custom ideas, characters and its corresponding fuzzy values by introducing the fuzziness theory into matter-element anal- ysis, and gave an extensive formalized expression of customer ideas in the fuzzy matter-element model, furthermore, used the fuzzy matter element optimization method to transform the multi-object optimization problems to the single-object optimization problems. Then, it gave the solution procedure of adaptive mutation particle swarm optimization algorithm （AMPSO） and compared it with that of GA to show the validity of this algorithm. Finally, applied the algorithm into the mining of customer idea in ear＇ s appearance feature, which effectively guided the implementation of product innovation and greatly improved the front end performance.