摘要： 利用约束集推移变换和前置矩阵降阶有效解决了软件过程活动规划中多约束条件类问题，提出了一种面向微观层次以活动为主体的软件过程活动规划和资源配置方法。通过染色体的矩阵编码形式结合相对顺序交叉算子与二级变异的操作方法对其进行了优化，实例证明了该算法的可行性和有效性。

关键词: 软件过程, 遗传算法, 活动规划, 资源配置

Abstract: Recursive transform on constrained set and order reduce with preceding matrix are used to deal with multi-constrained conditions in activities schedule of software process. A method based on genetic algorithm, of which the individuals’ structure is matrix, for activities schedule and resource assignment is proposed. Relative sequential crossover operator and mutation operation of two layers are adopted. The results show the method is effective to cut costs and compress duration of software process.

Key words: Software process, Genetic algorithm, Activities schedule, Resource assignment