串并行软件测试资源动态分配方法仿真

Simulation of Dynamic Allocation Method For Serial Parallel Software Test Resources

针对传统的软件测试资源分配方法在分配过程中产生网络延时导致资源分配时间长,效率低以及资源利用率低等问题,利用一种基于BPSO算法的串并行软件测试资源进行动态分配方法,分析串并行软件测试资源、测试代价以及软件可信度间的规则联系,构建了以测试资源为约束,以最高可信度以及最小化分配成本为目标的多目标动态分配模型,采用BPSO算法对多目标动态分配模型进行求解,在种群适应度函数中添加惩罚函数,将多目标动态优化问题转变为单目标动态优化问题;通过更新种群中粒子的位置、速度得到群体最优值,将全局最优解视为测试资源动态分配结果。实验结果表明,所提分配方法与传统分配方法相比,分配后的串并行软件测试资源利用率更高,分配用时更短、效率更高。

Traditionally,the allocation method of software test resource often causes problem of network delay,leading to long resource allocation time,low efficiency and the waste of resource.Therefore,this paper puts forward a method for dynamically distributing serial-parallel software test resources based on BPSO algorithm.Firstly,the rule relation between serial-parallel software test resources,test costs and software credibility was analyzed.In order to maximize credibility and minimize allocation costs,the multi-objective dynamic allocation model that took the test re⁃sources as the constraint was built.Moreover,BPSO algorithm was used to solve multi-objective dynamic allocation model.Then,the penalty function was added to the population fitness function,the problem about multi-objective dy⁃namic optimization was transformed into a problem about single-objective dynamic optimization.By updating the posi⁃tion and speed of particle in population,the corresponding optimal value was obtained.Finally,the global optimal so⁃lution was regarded as the result of dynamic allocation of test resource.Simulation results show that,compared with the traditional allocation method,the resource utilization ratio of serial-parallel software test after the allocation of proposed method is higher.Meanwhile,the allocation time is shorter and the efficiency is higher.