Abstract: Current methods usually take long retrieval time and low accuracy when retrieving components to satisfy user’s complex requirements. This paper presents a novel component retrieval method based on simulated annealing algorithm to quickly and iteratively select approximate optimal set of components. The realization of this method includes the definition of serial matching solution and its neighborhood, the formula of probability of acceptance, as well as the process of iteratively selecting sets of components with the lowest cost. Experiments show positive results in terms of execution time and retrieval accuracy.

Key words: software reuse, component retrieval, functional requirement, active services

摘要： 现有构件检索方法在面对用户复杂功能需求时检索时间较长而且检索结果不够准确，针对该问题，提出基于模拟退火算法的构件检索方法，定义构件检索过程中的有序匹配解及其邻域，并结合构件的分类描述特征设计模拟退火指标函数，在此基础上迭代搜索出覆盖用户功能需求且成本近似最低的构件集合。实验比较结果证明，基于模拟退火算法的构件检索方法在检索速度和准确度方面有一定的提高。

关键词: 软件复用, 构件检索, 功能需求, 主动服务

CLC Number: 

• TP311