基于改进离散乌鸦搜索算法的测试优选方法

Test optimization method based on improved discrete crow search algorithm

针对现有测试优选方法存在全局寻优能力不足的问题,提出了一种基于改进离散乌鸦搜索算法(IDCSA)的测试优选方法。首先,对经典乌鸦搜索算法(CSA)进行了离散编码,提出了一种自适应感知概率(AP)调整策略,并改进了CSA的乌鸦位置更新公式。其次,通过“超外差接收机”测试优选实例对IDCSA的优化性能进行了仿真验证,在多次重复仿真后发现,IDCSA能以接近100%的概率得出目前已知的总测试代价为6的全局最优解,所需平均迭代次数约为75次。验证结果表明,与现有的一些基于元启发式算法的测试优选方法相比,文中所提出的基于IDCSA的测试优选方法能更好地避免陷入局部最优,从而能够更稳定地得出测试优选问题的全局最优解。最后,将该方法应用于某型号独立型有源电子式电流、电压互感器(ECVT)的测试优选,从而验证了算法改进措施的有效性。该研究将为乌鸦搜索算法在测试优选领域的进一步应用提供基础。

Aiming at the problem that the existing test optimization methods have insufficient global optimization ability,this paper presents a optimization method of test selection based on the improved discrete crow search algorithm(IDCSA).Firstly,the classical crow search algorithm(CSA) was discretely coded,an adaptive awareness probability(AP) adjustment strategy was proposed,and the crow position updating formula of CSA was improved.Secondly,the performance of IDCSA was simulated and verified by the test optimization example of “superheterodyne receiver”.After repeated simulations,it was found that IDCSA can get the global optimal solution with the known total test cost of 6 with a probability of nearly 100%,and the average number of iterations required was about 75.The result shows that compared with some existing test optimization methods based on meta heuristic algorithm,the test optimization method based on IDCSA proposed in this paper can better avoid falling into local optimization,and thus can more stably get the global optimal solution of the test optimization problem.Finally,the method was applied to the test optimization of an independent active electronic current and voltage transformer(ECVT),and the effectiveness of algorithm improvement measures was verified.This research will provide a basis for the further application of crow search algorithm in the field of test optimization.