项目名称: 基于遗传算法的智能化的动态服务选择研究
项目编号: No.60872042
项目类型: 面上项目
立项/批准年度: 2009
项目学科: 金属学与金属工艺
项目作者: 张成文
作者单位: 北京邮电大学
项目金额: 30万元
中文摘要: 针对目前面向QoS限制的服务选择技术中存在的可扩展性差、不具有动态适应性、实时处理差、稳定性差等问题,采用遗传算法与其他理论或算法相结合的方法,通过理论分析和仿真验证相结合的手段,探索与面向服务质量的服务选择相关的理论问题。 本项目取得了以下研究成果,包括可扩展的混合聚类QoS模型、可适配的编码方式、Petri网与遗传算法在服务组合方案选择上的结合技术、动态适应度评估策略、基于混沌算法的遗传算法收敛性、基于精英选择的进化策略、基于聚类排序选择的个体选择方法、减少服务选择算法运行时所带来的网络负载的方法、基于种群多样性控制的交叉和变异技术、种群规模动态适应技术、基于单纯形与遗传算法相结合的选择机制、基于云模型的不确定性QoS计算方法、基于服务使用者评价的信誉度度量方法、基于模糊逻辑和自适应粒子群优化算法的服务选择方法、基于服务网的多媒体会议服务编制正确性验证方法。 本项目的研究成果将会推进服务理论与服务应用的向前发展。
中文关键词: 服务选择;QoS;遗传算法;可扩展;适应性
英文摘要: Aiming at service selection algorithm with extendability, dynamic adaptability, good real-time characteristic and stability, we make the research about how a special genetic algorithm and other selection theories work well together to build a service selection algorithm characterized by fine stability, extendability, dynamic adaptability and practicability. The obtained research achievements include extendable hybrid QoS-clustering model, adaptable coding method, Petri Net and Genetic Algorithm for web service selection, dynamic fitness evaluation policy, convergence capability of genetic algorithm based on chaos theory,evolution policy based on elitist selection, individual selection based on cluster ranking, method to decrease network overload of the running of selection algorithm, crossover and variation based on population diversity control, dynamic adaption of population size, selection based on genetic algorithm and Simplex Method, QoS uncertainty computation based on cloud model, reputation evaluation approach for web service selection, service selection approach based on fuzzy logic and adaptive particle swarm optimization algorithm, correctness verification approach for multimedia conferencing service based on service net. All of research achievements will promote the theory and application of service computing.
英文关键词: Service Selection; QoS; Genetic Algorithm; Extendability; Adaptability