基于改进粒子群算法的多目标智能组卷研究  

Research on multi-object intelligent test paper generation based on improved pso algorithm

作　　者：张顺淼[1] 

机构地区：[1]福建工程学院计算机与信息科学系,福建福州350108

出　　处：《福建工程学院学报》2011年第1期86-91,共6页Journal of Fujian University of Technology

基　　金：福建工程学院科研基金资助项目(GY-Z0566)

摘　　要：组卷问题是一个多约束多目标组合优化问题。建立了一种新的组卷数学模型,提出了一种改进粒子群算法(IPSO)的智能组卷算法。在组卷之前,先采用贪婪算法进行预处理,有效提高了算法的收敛速度。在进化过程中引入保优策略,避免适应值高的粒子被淘汰。采用自适应交叉和变异算子进行调整,避免了局部收敛现象。实验结果表明,改进的粒子群算法应用在多目标智能组卷中具有较好的组卷性能。Test paper generation is a multi-constraint and multi-objective optimization issue.A new mathematical model of intelligence test generation was established.An improved particle swarm optimization（IPSO） algorithm was presented for intelligent test paper generation.To effectively improve the convergence speed,pretreatment with greedy algorithm was adopted.Strategy for ensuring qual-ity during evolution was introduced to avoid the elimination of particles with high fitness value.To avoid the local convergence,adaptive crossover and mutation operator were adopted.The results show that the improved pso algorithm employed in multi-object intelligent test paper generation is efficient.

关 键 词：多目标智能组卷 数学模型 改进粒子群算法 

分 类 号：TP317[自动化与计算机技术—计算机软件与理论]