智能优化算法性能定量评价方案  

Quantitative Evaluation Scheme of Intelligent Optimization Algorithm Performance

作　　者：蓝永康[1] LAN Yongkang(The Department of Basic,Xi'an Siyuan University,Xi'an Shaanxi 710038,China)

机构地区：[1]西安思源学院基础部,陕西西安710038

出　　处：《信息与电脑》2022年第22期92-95,共4页Information & Computer

摘　　要：文章定量量化了算法的精度、时间复杂度、收敛效率和稳定性4个方面,给出一套评价智能优化算法性能的方案。首先,设计10个基准测试函数,对算法的全局搜索能力、局部搜索能力和搜索效率进行定量测试。其次,对比随机搜索算法的测试结果,将其转化为客观的定量分数。最后,应用该评价方案测试遗传算法、差分进化算法、粒子群算法、模拟退火算法、人工蜂群算法、万有引力算法、鲸鱼算法以及布谷鸟算法,并根据测试结果分析各算法的性能和优缺点。该评分方案可为算法的开发和应用提供极大便利。A scheme to evaluate the performance of intelligent optimization algorithm is presented,and the algorithm is evaluated quantitatively in the four aspects:precision,time complexity,convergence efficiency and stability.Firstly,10 benchmark functions are designed to quantitatively test the global search ability,local search ability and search efficiency of the algorithm.Secondly,by comparing with the test results of the random search algorithm,the results are converted into objective quantitative scores.Finally,the evaluation scheme was applied to test the genetic algorithm,differential evolution algorithm,particle swarm optimization algorithm,simulated annealing algorithm,artificial bee colony algorithm,universal gravitation algorithm,whale algorithm and cuckoo algorithm,and the performance,advantages and disadvantages of each algorithm were analyzed according to the test results.This scoring scheme provides great convenience for the development and application of the algorithm.

关 键 词：智能优化算法 元启发算法 随机搜索算法 基准测试函数 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]