基于改进粒子群算法的变异体选择优化

变异测试是常用的测试方法之一，变异测试分析的过程中计算开销会比较大，问题主要集中于测试过程中会产生大量的变异体，为了减少变异体的数量，提出用标准粒子群聚类算法进行选择优化，但标准粒子群算法在被测数据量增加到一定数量的时候，它的迭代次数就会增加、收敛速度就会下降。针对以上问题提出基于改进的粒子群算法对变异体进行选择优化。通过对变异体集合进行聚类分区，增强变异体集合的多态性，从而对粒子群算法改进优化。实验结果表明在不影响测试充分度的前提下，使变异体的数量大幅度减少，同时与K-means算法以及标准粒子群算法相比之下，改进后的方法具有更好的优化效果。

Mutants selection based on improved particle swarm optimization algorithm

Mutation test is one of the most common testing methods. In the process of mutation test analysis, computational overhead will be relatively large. The problem is mainly that mutation test will generate a lot of variants in that process,therefore, in order to reduce the number of variants,a standard particle swarm optimization algorithm is proposed to select and optimal the number of the variants. However, the Standard Particle Swarm Optimization will have large iteration times and slow convergence velocity if the measured data reach a certain amount. To solve the above issues, the paper provides mutants selection and optimization which based on improved particle swarm algorithm. By clustering and partitioning the test data set and enhancing the polymorphism of the mutations set, it optimizes the particle swam algorithm. Experimental results show that under the premise of not affecting the adequacy of the test , and greatly reducing the number of variants, the improved method has a better optimization results compare with the K-means algorithm and the standard particle swarm algorithm.