中文摘要：

为了提高智能组卷质量，提出一种基于小生境自适应遗传模拟退火算法的智能组卷策略．该算法动态调整交叉和变异概率进行遗传操作，对中间种群进行小生境选择和模拟退火操作，从而增强了种群多样性，有效克服了遗传算法局部收敛和“早熟”的缺点．文章针对各约束条件建立了组卷数学模型，给出了基于期望平均分的难度分布函数和小生境自适应遗传模拟退火组卷模型．大量测试数据表明，该方法是一种有效可行且实用的组卷方法．

英文摘要：

In order to improve the quality of test paper auto-generating, an intelligence algorithm based on niche adaptive genetic simulated annealing was proposed. It adjusted the crossover and mutation probability dynamically for genetic operation, and applied niche selection and simulated annealing operations to the intermediate populations, thus the variety of populations was advanced greatly. It overcame the shortcoming that GA is easy to fall into a local optimum and prematurity. The model of generating test paper with con- straints based on niche adaptive genetic simulated annealing algorithm was set up, where the difficulty distribution function based on the expectation of test score was given. Finally, the algorithm was proved effective, feasible and practical by testing data.