基于自然数编码遗传算法的并行测试技术

针对并行测试任务调度复杂、难以优化的问题，提出了一种基于自然数遗传算法的任务调度优化算法。该算法将并行测试任务调度转化为对串行测试任务序列的搜索，为了提高搜索效率，引进自然数编码遗传算法搜索最优解或近似最优解，将串行测试序列作为染色体，进行选择、交叉、变异等操作。通过仿真证明该算法能够迅速得到并行测试任务调度的最优解或近似最优解。对自然数编码遗传算法的各个算子对模式的作用进行了理论分析，并得到了对应的模式定理，从理论上说明了该方法的合理性。

Parallel test using natural coding genetic algorithm

Considering parallel test task scheduling in automatic test system is complicated and difficult to optimize, a task scheduling arithmetic based on natural coding genetic algorithm is proposed. This arithmetic transforms searching parallel test tasks sequences into searching serial test tasks sequences. The natural coding genetic algorithm, which could improve efficiency, is applied to search for optimum or near optimum sequences. The serial test tasks sequences are regarded as chromosomes, which are selected, crossed, and mutated. According to the simulating results, using this method the optimum or approximate optimum result could be gain quickly. The effect of the operators is analyzed, and then the schema theorem is presented, which proves the rationality of genetic algorithm used.