智慧教育云平台中多约束分配问题的算法设计与研究

发布时间：2018-03-07 15:41

本文选题：智慧教育　切入点：多约束分配　出处：《南昌大学》2016年硕士论文　论文类型：学位论文

【摘要】：中华五千年历史中,教育一直都是国家和人民关注的热点,随着时间的推移,社会的变迁,教育形式也在一步一步地改变。短短几十年时间,信息化技术发展突飞猛进,数字化教育还没来得及普遍推广,基于大数据、物联网、云计算、人工智能的智慧教育云平台已经开始崭露头角了。为了让教育工作者把更多的时间与精力花在教学理念和教学方法的提升上,我们有必要将一些诸如新生分班、教师排课、考试安排、试卷切题统分、成绩分析、学习能力评估等复杂繁琐却需要人工操作的工作,通过计算机辅助,使其智能化完成。经过调查分析,发现新生分班、教师排课、考试安排这类多约束分配、NP难问题是当前学校迫切想解决的棘手问题。针对这三个问题,本文详细分析了中小学与高校各自不同的要求,总结出必须满足的硬约束条件和可优化的软约束条件。为了找到解决这三个问题的最佳算法,本文对多约束分配问题的常用算法进行了分析比较,阐述了其基本原理和适用范围。由于分班问题中新生分数的规律性、性别的特殊性、生源地区域性等特征,本文提出了基于回溯的新生分班算法;而排课与考试都是将课程、教师、教室、班级与时间关联起来,并且两者约束条件具有一定相似性,因此将二者归纳成时间表问题。基于时间表问题的复杂性,结合遗传算法的全局解搜索能力和蚁群算法的并行反馈机制,最后确定了基于蚁群遗传的时间表算法。文中提出的新生分班算法及时间表算法,已实际应用到“智慧教育云平台”中,实际应用结果表明该算法具有较好的鲁棒性,有效地解决了“智慧教育云平台”中的若干难点问题。
[Abstract]:In the five thousand years of China's history, education has always been the focus of attention of the state and the people. With the passage of time, social changes and educational forms have changed step by step. In a short period of several decades, information technology has developed by leaps and bounds. Digital education has not yet been widely promoted, based on big data, the Internet of things, cloud computing, The intelligent education cloud platform of artificial intelligence has begun to emerge. In order for educators to spend more time and energy on the promotion of teaching ideas and methods, it is necessary for us to divide new students into classes and teachers to schedule classes. The complicated and complicated work, such as examination arrangement, test paper grading, score analysis, learning ability evaluation, which needs manual operation, is accomplished intelligently through computer aid. Through investigation and analysis, it is found that new students are divided into classes and teachers arrange classes. The NP-hard problem of multi-constraint assignment is the thorny problem that schools are eager to solve at present. In view of these three problems, this paper analyzes in detail the different requirements of primary and secondary schools and colleges and universities. In order to find the best algorithm to solve these three problems, the common algorithms of multi-constraint assignment problem are analyzed and compared. This paper expounds its basic principle and application scope. Because of the regularity of freshmen's scores, the particularity of gender and the region of students' origin, this paper puts forward a new class dividing algorithm based on backtracking, and the course scheduling and examination are both general courses. Teachers, classrooms, classes and time are associated with each other, and their constraints are similar, so they are reduced to timetable problems. Combined with the global search ability of genetic algorithm and the parallel feedback mechanism of ant colony algorithm, the algorithm based on ant colony genetic algorithm is finally determined. It has been applied to the "Intelligent Education Cloud platform". The practical application results show that the algorithm is robust and effectively solves some difficult problems in the "Intelligent Education Cloud platform".
【学位授予单位】：南昌大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18;TP393.09

【参考文献】