基于混合人工蜂群算法的并行测试任务优化研究

Research on Parallel Test Task Optimization Based on Hybrid Artificial Bee Colony Algorithm

摘要:

并行测试技术可以同时进行多个任务的测试，提高资源利用率，节约测试成本；并行测试调度问题是一种复杂的组合优化问题，是并行测试技术的核心要素；并行测试系统作为并行测试技术的载体，自身的性能和求解效率尤其重要；对并行测试完成时间极限定理进行了研究，建立了并行测试任务调度的数学模型，分析了传统元启发式算法求解并行测试问题的不足，提出了基于动态规划的递归搜索技术和人工蜂群算法相结合的混合人工蜂群算法，并采用整数规划精确算法和遗传算法对混合人工蜂群算法进行验证；得出结论采用混合人工蜂群算法进行并行测试任务的调度节约了接近50%的时间，降低了约20%的硬件资源占用，提高了测试效率，可以满足工程实际的应用。

Abstract:

Parallel testing technology can simultaneously test multiple tasks, improve resource utilization, and save testing costs; Parallel test scheduling problem is a complex combinatorial optimization problem, and it is the core element of parallel test technology; As the carrier of parallel testing technology, the performance and solving efficiency of parallel testing systems are particularly important; The limit theorem of parallel test completion time is studied, the mathematical model of parallel test task scheduling is established, and the shortcomings of traditional meta heuristic algorithms for solving parallel test problems are analyzed. A hybrid artificial bee colony algorithm based on the combination of recursive search technology of dynamic programming and artificial bee colony algorithm is proposed, and the hybrid artificial bee colony algorithm is verified by using the precise algorithm of Integer programming and genetic algorithm; The conclusion is that using hybrid artificial bee colony algorithm for scheduling parallel testing tasks saves nearly 50% of time, reduces about 20% of hardware resource usage, improves testing efficiency, and can meet practical engineering applications.