基于线性递减系数粒子群优化算法的组卷实现

白雁

摘  要： 为了避免目前常用的组卷算法组卷时间长、程序结构复杂、收敛速度慢等缺陷，提出基于线性递减系数粒子群优化算法的组卷策略。通过调整惯性系数，使得步长较小，惯性权系数的变化幅度小，这种减小趋势较为缓慢的方法能够避免陷入局部最优。并对数学模型以及线性递减惯性权系数进行了理论设计，同时通过编程实现了该算法。测试结果表明加入线性递减系数后运算迭代次数明显减少，证明加入线性递减系数后的组卷策略收敛性好，能够高效准确地按照一定的预期条件进行组卷，符合预期要求。

关键词： 组卷; 粒子群优化算法; 线性递减惯性权系数; 适应度函数

中图分类号： TN911?34                    文献标识码： A                            文章编号： 1004?373X（2014）24?0041?04

Implementation of test paper generation based on particle swarm optimization algorithm with linear decreasing inertia weight

BAI Yan

（The Distance Education School， Xian Jiaotong University， Xian 710048， China）

Abstract：In order to avoid defects in the commonly used test paper generation algorithm， such as too much time taking， complicated program structure and low velocity of convergence， a test paper generation strategy of  particle swarm optimization algorithm based on linear decreasing inertia weight is proposed. The step size becomes smaller and the inertia weight changes less by adjusting the inertia coefficient. The relatively slow decreasing trend method can avoid falling into local optimum. The theoretical design for mathematical model， linear decreasing inertia weight was carried out. The algorithm was realized by programming. Test results show that the addition of linear decreasing coefficient can greatly reduce the iteration times， can make the convergence characteristic better， and can efficiently and accurately generate test paper according to the expected conditions.

Keywords： test paper generation; particle swarm optimization algorithm; linear decreasing inertia weight; fitness function

0  引  言

教育信息化促进了教育领域的全面变革，其中传统的考试方式也慢慢开始逐步被在线考试这一形式所取代。如何快捷、有效、科学地组卷是在线考试系统的技术核心。合理的组卷策略不仅仅是完成一份考题，而是要像人工考试那样能够按照总分，题型，章节，难度等信息筛选相应的试题组成具有一定难度和区分度的试卷，这样能够更加客观，公平，真实地反映出学生对知识点掌握的情况。

组卷算法实际上是多目标约束优化问题，合理选择组卷算法是解决问题的关键。目前常用的组卷算法有随机组卷法、回溯法、遗传算法、鱼群算法、粒子群优化算法等。但这些算法存在着组卷时间长，程序结构复杂，收敛速度慢，计算复杂等缺陷[1]。……