【摘要】 将服务组合的问题视为一个Agent动作规划的问题,给出了一个基于动作规划的服务组合模型。进而结合遗传规划的可变分层结构,提出了一个适应于上述模型的服务匹配算法。仿真实验表明,在具有关联特征的服务选择空间中,该算法比其他算法具有更好的寻优能力和更快的速度。更多还原

【Abstract】 Based on the feature abstraction in semantic service,the elementary service can be abstracted as action rule unit(ARU).So,a services composition model based on action planning was proposed.In order to satisfy the functional and nonfunctional needs,a service matching algorithm based on genetic programming(GP-SMA),was introduced by integrating with the variable and layered structure of genetic programming.The action rules among these ARU guide the planning process.Simulation results show that this algorithm exhibits some good performances such as find-best ability and efficiency in the selection space of associated service.更多还原