Solving university course scheduling problem using genetic algorithm and analyzing results with other algorithms
Abstract
A study on course timetabling problem which is a combinatorial optimization NP-hard problem. The aim of this thesis is to find optimal or near optimal solution of course scheduling for Computer Science and Engineering Department of BRAC University. Different solution methods for course timetabling exist hence in this thesis Genetic Algorithms is used to generate feasible solution and Q-learning is action for evaluating results. Experimental data sets are parsed from a given structure. Different constraints are handled with discrete fitness evaluation. Schedule conflicts are handled after producing random generation. Finally, results are tested
according to their performance and presented with a feasible representation mode.