【作者】 吴艳娟；

【导师】 王成良；

【作者基本信息】 重庆大学 ， 计算机软件与理论， 2011， 硕士

【摘要】 将多个工作流服务聚合为具有特定功能的服务来满足用户对复杂功能的需求已经成为一个研究热点。由于工作流服务不断增加,服务聚合往往会出现大量的备选方案,用户期望从这些方案中选择满足Qos全局最优的工作流服务聚合流程。现有服务聚合方法大多都是基于Qos局部最优原则,无法满足对Qos全局最优的需求。本文将满足Qos全局最优的工作流动态服务聚合问题转化为带约束的多目标优化问题。针对粒子群算法求解多目标优化问题上的优势,提出一种改进的混合多目标粒子群算法(IHMOPSO)。算法引入遗传算法中的交叉变异策略,并通过自适应的惯性权重调节和基于拥挤距离的全局最优解概率选择机制,改善了多目标粒子群算法收敛慢、容易陷入局部最优的缺陷。本文主要工作包括:①在对工作流动态服务聚合问题研究基础上,将满足Qos全局最优要求的动态服务聚合问题转化为带约束的多目标优化问题。②通过对多目标粒子群优化算法中的几种关键理论的分析,针对多目标粒子群算法的主要问题,提出一种改进的混合多目标粒子群优化算法。该算法利用遗传算法中的交叉变异策略,对精英种群中个体进行交叉变异,同时采用基于拥挤距离的全局最优解概率选择机制,保证Pareto最优集的多样性;自适应的惯性权重的设置,保证算法在全局搜索和局部搜索之间达到平衡;将种群划分为精英种群和普通种群,保证算法的收敛速度。③构建基于Qos的工作动态流服务聚合多目标优化模型,采用改进的混合多目标粒子群优化算法求解该多目标优化问题。④对本文所提方法进行实验验证:结合祥弘办公自动化系统的项目,构建工作流服务聚合实例模型,采用IHMOPSO算法对工作流服务聚合多目标优化问题进行求解。对算法的收敛速度及解集分布进行分析,说明本算法的可行性,将实验结果与同类方法比较,验证本算法的有效性。通过对本课题实验结果进行分析,本算法可收敛到一组满足Qos全局最优的服务聚合流程供用户选择,实验结果表明本算法具有较好的收敛速度和种群多样性。更多还原

【Abstract】 It has become a hot topic to aggregate multiple workflow services into one with complex function to meet the needs of users. As a result of the number of workflow service increasing, the service aggregation often has a lot of options, but users expect to get the workflow service process meeting the Qos global optimum. However, most existing service aggregates are based on local principle, so it can not meet the needs of the Qos global optimum.The main purpose of this paper is to transform the workflow dynamic service aggregation with Qos global optimum into multi-objective optimization problem with constrained. According to the advantages of solving multi-objective optimization problems by particle swarm optimization, this paper proposes a hybrid multi-objective particle swarm optimization (IHMOPSO). The algorithm, which includes the crossover and mutation strategies from genetic algorithm and selects mechanism which is through the adaptive inertia weight regulation and the probability of global optimum based on crowding distance, improves the defect on the slow convergence and easily falling into local optimal of the multi-objective particle swarm optimization.The main contents of this paper can be summarized as follows:①Transform the Qos global optimum dynamic service aggregation into multi-objective optimization problem with constrained on the basis of studying the workflow Dynamic Service Aggregation.②Through the analysis of critical theory on multi-objective particle swarm optimization, to solve the main problem, an improved hybrid multi-objective particle swarm optimization algorithm is proposed, which uses the strategy of crossover and mutation in genetic algorithm to cross and mutate the individuals in the elite population, which adopts the probability selection mechanism of global optimum based on crowding distance to ensure the diversity of Pareto optimal set, which sets adaptive weight to ensure the balance between global search and local search, which divides population into the elite and general ones to ensure the convergence rate.③Constructs the multi-objective optimization model of workflow service aggregate with Qos, and adopts the improved hybrid multi-objective particle swarm optimization algorithm to solve the multi-objective optimization problem.④This paper constructs a model of work-flow service aggregation combined with the office automation system of XiangHong, and adopts IHMOPSO to solve the multi-objective optimization problem of work-flow service aggregate. It proved to be feasibility with the analysis of convergence speed and the distribution of solution set, and validity compared the experimental results with similar methods.Trough analyzing the result of the project, it proves the algorithm can converge to a set of the aggregation process meeting Qos global optimum, and has better convergence speed and population diversity.更多还原