中文摘要：

为了充分体现服务质量（QoS）的不确定性和用户偏好的模糊性,本文将模糊集理论引入基于QoS的Web服务组合中,将不适合精确表示的QoS属性和用户偏好等信息用三角模糊数表示.然后基于权重和法计算模糊总目标,通过设计新的模糊数比较方法,改写Pareto支配关系,将基于模糊数比较的单目标优化问题转化为多目标优化问题,并设计模糊多目标遗传算法（FMOGA）求得Pareto最优解集.该方法不仅能够得到更加贴近实际情况的优化解,同时也解决了多属性决策方法无法对大量候选服务进行全局优化的问题.最后通过实验验证了该算法的有效性和优越性.

英文摘要：

In order to fully reflect the uncertainty of quality of service（QoS） and the ambiguity of user preferences,this article introduces the fuzzy sets theory into QoS-based Web services composition,the QoS attributes and user preferences that are not suitable for accurate representation are described by triangular fuzzy number.Then the fuzzy total goal is calculated based on weighted-sum approach,and the pareto dominance relationship is redesigned based on the new approach of the fuzzy number comparison,the single objective optimization problem based on fuzzy numbers comparison is converted to multi-objective optimization problem,and the fuzzy multi-objective genetic algorithm（FMOGA） is designed to obtain pareto optimal solution set.This method can not only obtain the optimal solutions which are closer to the actual situation,but also solve the question that the multiple attribute decision making methods can not global optimize numerous candidate services.At last,experiments demonstrate the effectiveness and advantage of the method.