改进的量子遗传算法及其在测试数据生成中的应用

摘要

This paper proposed an Improved Quantum Genetic Algorithm (IQGA) for the problem of slow convergence in test data generation. There are two main improvements. First, every bit of every individual was reversed to conduct the evolution; second, the binary individuals were mutated after measurement, instead of the traditional exchange of the probability amplitude of quantum bits. IQGA was applied into test data generation. The experiments on three basic programs prove that IQGA is better than QGA in terms of coverage rate and the number of iterations. IQGA can not only ensure the right direction of the evolution of populations, but also avoid premature phenomenon, and it can get the solution at a faster convergence speed.%针对测试数据自动生成中收敛速度不够快的缺点,提出一种改进的量子遗传算法( IQGA),其对量子遗传算法的主要改进是:1)在个体更新时,对个体的某一位取反,将取反后的个体用于指导下一代个体的进化；2)对测量后的二进制个体进行变异,而不是传统的互换量子比特的概率幅.将IQGA用于测试数据生成,通过对三个基础程序进行实验,结果表明IQGA在覆盖率和迭代次数两个方面都优于传统量子遗传算法.IQGA不仅能保证种群朝着正确的方向进化,同时有效地避免了早熟现象,能以更快的速度搜索到目标解.