摘 要 高校教务管理部门在整个教学过程中起着组织、协调以及服务的作用，其中排课是最基础，也是最繁琐的一项任务。由于各高校前期实行扩招政策，目前在校学生数平均达到2万人左右，面对新形势，在教学资源有限的情况下采用传统的人工排课，不仅会使排课压力增大，而且会因为约束条件的叠加，导致排课问题复杂化，所以采用智能排课系统替代传统排课模式是很有必要的。当前排课系统中普遍单一使用遗传算法、蚁群算法、模拟退火算法等算法求解排课问题，求解时有耗时长、易陷入局部最优解等缺陷，因此本文提出一种基于遗传算法和蚁群算法混合的方法求解排课问题。 首先，确定了排课过程相关的因素与约束条件，在此基础上建立排课问题的数学模型和衡量课程表优劣的适应度函数，基于以上分析和对排课业务流程的梳理，本排课系统包含信息录入、教务管理、智能排课等模块。 其次，在系统中将遗传算法和蚁群算法分别引入排课问题。对遗传算法中染色体编码、冲突检测与遗传算子的操作方式进行设计。在蚁群算法的基础上构建排课问题的二分图模型，并对蚁群算法的不足进行讨论，采用最大-最小蚂蚁系统和动态启发函数对蚁群算法进行改进。为了充分发挥遗传算法和蚁群算法各自的优势，同时克服遗传算法后期无法充分利用反馈信息和蚁群算法初期搜索慢的缺陷，本文将遗传算法和蚁群算法混合的方法应用于排课问题中，结合遗传算法前期搜索效率高和蚁群算法搜索后期能够快速获取最优解的优点，共同解决排课问题，并进行Matlab仿真实验验证混合算法是可行的，仿真结果表明混合算法可解决单一算法缺陷的问题。 最后，本文应用遗传-蚁群混合算法在Java平台完成智能排课模块，并基于Spring框架实现排课系统。

ABSTRACT The administrative management department of colleges and universities plays the role of organization, coordination and service throughout the teaching process. Among them, course-scheduling is the most basic and the most cumbersome task. Due to the expansion of the colleges and universities in the early stage, the current number of students in the school averages about 20,000. In the face of the new situation, the use of traditional manual courses in the case of limited teaching resources will not only increase the pressure on class scheduling, because of the superposition of constraints, the course-scheduling problem is complicated, so it is necessary to adopt the intelligent course-scheduling system instead of the traditional course-scheduling mode. In the current course-scheduling system, Genetic Algorithm, Ant Colony Algorithm, Simulated Annealing Algorithm and other Algorithms are commonly used to solve the course-scheduling problem. The solution is time-consuming and easy to fall into the local optimal solution. Therefore, this paper proposes a hybrid method based on Genetic Algorithm and Ant Colony Algorithm to solve the course-scheduling problem. Firstly, the factors and constraints related to the course-scheduling process are determined. On this basis, the mathematical model of the course-scheduling problem and the fitness function of measuring the merits of the curriculum are established. Based on the above analysis and the sorting of the business process of the course-scheduling, the course scheduling system includes modules for information entry, educational administration, and intelligent scheduling. Secondly, the Genetic Algorithm and the Ant Colony Algorithm are introduced into the course-scheduling problem respectively in the system. The operation of chromosome coding, collision detection and genetic operators in Genetic Algorithm is designed. Based on the Ant Colony Algorithm, the bipartite graph model of the course-scheduling problem is constructed, and discussed the deficiencies of the Ant Colony Algorithm. The Ant Colony Algorithm is improved by the Max-Min Ant System and the dynamic heuristic function. In order to give full play to the advantages of Genetic Algorithm and Ant Colony Algorithm, and overcome the shortcomings of the late use of feedback information and Ant Colony Algorithm in the late stage of genetic algorithm, in order to give full play to the advantages of Genetic Algorithm and Ant Colony Algorithm, at the same time overcome the late stage of Aenetic Algorithm can not make full use of feedback information and Ant Colony Algorithm early search slow defects,this paper applies the hybrid method of Genetic Algorithm and Ant Colony Algorithm to the course-scheduling problem. Combining the advantages of high search efficiency in the early stage of Genetic Algorithm and the ability to quickly obtain the optimal solution in the late stage of Ant Colony Algorithm, the problem of course-scheduling is solved together.It is feasible to verify the hybrid algorithm by Matlab simulation experiment. The simulation results show that the hybrid algorithm can solve the problem of single algorithm defect. Finally, this paper applies the GA-ACA hybrid algorithm to complete the intelligent course scheduling module on the Java platform, and implements the course-scheduling system based on the Spring framework.