CAJ | 학술논문

为克服经典遗传算法求解智能组卷时存在的盲目搜索、收敛速度慢和个体易"早熟"等缺点,文章提出了基于动态自适应技术遗传算法.利用求解问题的特征改进经典遗传算法的多个关键部分,采用动态自适应技术,提高算法的寻优速度.实验结果表明,采用改进式的自适应遗传算法实现的组卷策略具有收敛速度快、搜索精度高、鲁棒性强等特点,具有很好的性能与实用性.
위극복경전유전산법구해지능조권시존재적맹목수색、수렴속도만화개체역"조숙"등결점,문장제출료기우동태자괄응기술유전산법.이용구해문제적특정개진경전유전산법적다개관건부분,채용동태자괄응기술,제고산법적심우속도.실험결과표명,채용개진식적자괄응유전산법실현적조권책략구유수렴속도쾌、수색정도고、로봉성강등특점,구유흔호적성능여실용성.
In order to overcome the shortcomings of blind searching, the slow convergence rate and individual's easily precocious when using the classics genetic algorithm solution for the intelligent test paper composition, the article proposes genetic algorithm which is based on the dynamic auto-adapted technology, that is to use the characteristic of solution questions to improve many essential parts of the classic genetic algorithm and to use the dynamic auto-adapted technology to enhance the algorithm' s speed for searching. The experimental result indicates that the group volume strategy based on the auto-adapted genetic algorithm have the characteristics such as fast convergence rate, high search precision and strong robustness, and can achieve good usability.