Timetabling Problem(TP) is a Multiobjective Combination OptimizationProblem with constraints, and also has been proved NP-Completed.Genetic Algorithm(GA) is a high-effective randomly searching algorithm,based on the nature evolution. It is a very effective algorithm to solveNP-Completed Combination Optimization Problem.Genetic Algorithm is adopted to solve TP in this paper. The main contentsare as following:(1)This paper systematically and completely discusses factors, restrictionsand objectives attached to TP, and describes TP by mathematic model.(2)This paper improves the generalized coding method of GA, syntheticallyadopts three-dimension coding and self-adaptive designing method of crossoverand mutation probability, and brings forward a self-adaptive GA based onthree-dimension coding.(3)A course scheduling system based on above-mentioned improved GA hasbeen designed and realized by using Delphi7.0 as developing tool and MS SQLSERVER 2000 as database. And a TP example with 110 classrooms, 389schoolteaching events has been tested in objective space consists of three factorsof slot-superior, distributing-uniformity-on-day, and gym lessons time. The resultis satisfing.The system has been applied to one college in my province and improvescurriculum scheduling efficiency greatly. |