基于粒子群优化算法的智能抽题策略研究  

Research on the Tactic of Intelligent Test-Paper Generation Based on PSO Algorithm

作　　者：阎峰 安晓东 

机构地区：[1]山西省人事考试中心,山西太原030006 [2]山西人事信息中心,山西太原030072

出　　处：《中北大学学报（自然科学版）》2008年第4期333-337,共5页Journal of North University of China(Natural Science Edition)

基　　金：山西省自然科学基金资助项目(20060110035)

摘　　要：通过对智能抽题策略研究,建立了智能抽题数学模型,提出用于评价抽题效果的评价指标及其获取方法;在此基础上,考虑智能组卷过程对试卷质量的高要求以及优化参数的表示形式,文中将粒子群优化算法运用于智能抽题优化模型的求解中,取得了较好的效果.为证明该方法的有效性和优越性,与采用相同二进制编码方式的遗传算法进行比较,通过二者的适应度演变图可以表明粒子群算法(Particle SwarmOptimization,PSO)算法可以较早地达到目标的最优分配,在运行时间及抽卷优化结果上明显优于遗传算法(Genetic Algorithm,GA),能更好地适应智能抽题策略的要求.By studying the tactic of intelligent test-paper generation, the paper builds its mathematic model and puts forward the evaluation index and quantitative method for evaluating the effects of taking test questions. On the basis of it, computational algorithms for optimal design of the model of taking test questions are implemented by means of improved particle swarm optimization （PSO）. In order to show the effectiveness and superiority of this method, PSO algorithm is compared with genetic algorithm （GA） algorithm by adaptation evolvement chart of test-paper generation with the same binary encoding method. The result indicates that the PSO algorithm can achieve the superior object more quickly,the running time and optimization result of PSO surpass that of GA obviously. The method can satisfy well the strategy request of intelligent test paper generation.

关 键 词：智能组卷 数学模型 评价指标 粒子群算法 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]