一种基于遗传规划的服务匹配算法

A Services Matching Algorithm Based on Genetic Programming (GP)

当前,智能服务组合研究的重点主要集中在服务描述和服务匹配方面。文中首次将服务匹配视为一个状态描述含糊、操作定义不完备的规划问题,提出了一个基于遗传规划的服务匹配(Service Match Programming,SMP)算法和适应度评价函数。仿真实验表明,该算法在大尺度的服务选择空间中,引入了服务的关联特性,避免了局部最优现象的出现,具有更好的寻优能力和更快的速度。

Aim. It is well known that offering many kinds of limited services, distinct from each other, is far less efficient than offering all of them in an organized fashion for the user to choose freely from. This is exactly what we like to accomplish with our services matching algorithm. In the full paper, we explain our services matching algorithm in detail; in this abstract, we just add some pertinent remarks to the two topics of explanation： （1） services matching model and （2） services matching algorithm based on GP; the two subtopics of topic 1 are programming mapping （subtopic 1.1）, services matching model（subtopic 1. 2）; in subtopic 1.1,we point out that, based on the feature abstraction in service, the elementary service can be abstracted as programming unit and so, in subtopic 1.2, a services matching model based on action planning is proposed and the planning is driven by requirements; the four subtopics of topic 2 are the degree of services case comparability （subtopic 2.1）, the evaluate function of adaptability （subtopic 2.2）, the evolutionary operation （subtopic 2. 3） and the description of services matching algorithm based on GP （subtopic 2. 4）; in subtopic 2.4, we point out that, in order to satisfy the functional and nonfunctional needs, a services matching algorithm based on GP （SMP） is introduced by integrating with the variable and layered structure of GP. We performed simulation experiments, whose results are given in Figs. 1 and 2 in the full paper. These results show preliminarily that our SMP algorithm is indeed feasible and effective.