Design And Implementation Of Courses Arrangement System Based On Genetic Algorithms

Course arrangement is one of the most important and the most complicated tasks ofschool, and the fundamental essence of the school curriculum is set by a set of appropriateteaching time and space so that the whole educational work can be planned orderly.Recently, with the rapid development of the national vocational education, the great effortsof admissions policies make a substantial increase in the number of students．The originaluse of manual or artificial course scheduling has some difficulties, compared with thecomputer based course scheduling requirements.The genetic algorithm makes some references on biosphere mechanism of naturalselection and inheritance, and it uses the colony searching technology. It is a highlyparallel, self-adaptive, random and global seeking method. The genetic algorithm hashigher robustness, so it is especially suitable for large-scale, complicated and non-linearsystems. Usually, it demonstrates well than some other traditional seaching algorithms,which optimize the performance with more unique and more superior method.In this thesis, we use the genetic algorithm to solve course scheduling problem. Usingthe course table as chromosome objects for coding, this thesis schedules courses throughthree steps of genetic algorithm, including selection, crossover and mutation. As for thedesign of fitness function, the thesis considers some parameters, including the course time,course, fitness, class hour distribution of satisfaction, and the classroom teacher weeks. Asfor the distribution uniformity with genetic algorithm encoding, it is usually too long andfertility roulette, so it tends to cause the loss on good individual. We propose the solutionon population design and several rounds of roulette. This system also adopts the Browser/Server structure and the ASP.NET dynamic webpage technology, and it uses the SQLServer2005as the database management system, using the ADO.NET technology toconnect the web server and database server.The results show that this algorithm can solve the timetabling problem and can meetthe needs of the B/S structure online course scheduling. Moreover, this system has friendlyinterface and is easy to use. The experimental results validate the feasibility of theapproach.