The Research Of Course Scheduling System Based On Multi-Strategy Algorithm

The course scheduling in university is an extremely complex but very important thing, because it contains a lot of managements on information and data. According to investigation, with the deepening in the reform of university, the growth rates of student's number in enrollment expansion far exceed the teaching resources, which make the course scheduling more and more year by year. So it is urgent to develop an applicable intelligent scheduling system. So far, some universities and companies have developed software for course scheduling, and put them to use. The algorithms used in these systems are concentrated in some popular algorithms, for example, the genetic algorithm using artificial intelligence, the algorithm based on graph theory, and expert system, etc. But all of these algorithms have some problems, such as complexity of thinking, difficulties in development, and inexactness of optimization problems. So, except some advantages, there are some problems with various degrees in the software, such as low efficiency, poor versatility, lack of integrity, limitation of various conditions, and so on. One main reason causing these problems is using a single algorithm in these systems, so the course scheduling algorithm based on multi-strategy becomes an effective method to solve above problems.This paper analyses the applications of partition strategies, greedy algorithm, backtracking algorithm, genetic algorithm and taboos algorithm in course scheduling systems. These algorithms have their own advantages and disadvantages. To learn from others' strong points to offset one's weakness, it proposes a new idea about multi-strategy algorithm. The basic ideas of the algorithm describes as follow: first of all, according to the thoughts of partition strategies, the whole scheduling process was divided into classroom distribution and time distribution. Then, on the basis of the greedy algorithm, while distributing the classrooms, the system always chooses the classroom whose capacity is closest to the class and the type is suitable. While distributing the time, the system always chooses the best efficiency time for classes from all spare time under certain conditions. If deadlock occurs while distributing the time, it will back up the recent record and then rearrange to resolve conflicts. Combine the classroom and time. After traversing all the apply records, it will get the initial scheduling results, which basically meet the hard constraints of course scheduling. In order to make the schedule more human and the use of resources more reasonable, it uses the thoughts of genetic algorithm and taboo algorithm to optimize the initial results, and then it will get the final results.Based on the multi-strategy course scheduling algorithm, we design and develop the system with Microsoft Visual Studio 2008 system development tools, c# language and SQL Sever 2005 database development platform.With the design ideology of human-centered, in order to manage timetable information better and help management staff extricate from the heavy work in schedule arrangement, the system provides the functions of course prearrangement, auto arrangement, manual adjustment of syllabus, collision detection, syllabus save and so on, reach the purposes on reasonable allocation of time, full use of school resources, alleviation of working strength of the school course scheduling staff, optimization the learning process of students, and basically meet the demand of an intelligent course arrangement system.