自适应遗传算法在工程训练在线考试中的应用  

Application of adaptive genetic algorithm in engineering training on-line exam system

作　　者：朱婧[1] 戴青云[2] 王美林[1] 王森洪[1] 

机构地区：[1]广东工业大学信息工程学院,广州510006 [2]广东工业大学科技处,广州510006

出　　处：《计算机工程与应用》2013年第14期227-230,246,共5页Computer Engineering and Applications

基　　金：广东省教育厅产学研结合基地及科技成果转化重大项目(No.cgzhzd0608)

摘　　要：在工程训练中心车间信息化实现的基础上,针对工程训练管理系统中考试模块现有组卷方式所带来的抽重复题、组卷效率低下等问题,以满足在线考试的实时性要求。为此,给出一种改进的遗传算法,采用分段整数编码,改进初始种群的产生方法,有效提高了算法的收敛速度,并自适应调整遗传算子,在进化过程中增加去重题策略及最优个体保存机制,维护了种群多样性,保证了运算结果的质量。实验结果表明,该算法不但解决了系统组卷原有的问题,在迭代次数、运行时间和组卷精确度上均明显优于随机组卷法和简单遗传算法。On the basis of realizing the information technology in the engineering training center workshop, the management system of examination module has a lot of problems, such as forming repeated questions and the low efficiency of paper con- structing, in order to solve those problems to meet the real-time requirements of the on-line exam, this paper presents an improved genetic algorithm. In this algorithm, it adopts the subsection integer coding, improves the generation methods of the initial popu- lation, which can effectively improve the convergence speed. It also uses the adjustment method of adaptive genetic operator, then removes the repeated questions and adds best individual save mechanism in the evolutionary process, which can both ensure the diversity of population and acquire high quality. The experimental results show that the algorithm can not only solve the problem of the examination module, but also show the advantages over randomized algorithms and simple genetic algorithm in iterative times, running times and accuracy of composition test paper.

关 键 词：遗传算法 工程训练 智能组卷 整数编码 自适应 最优个体保存机制 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]