Solving Examination Timetabling Problem Using Crow Swarm Optimization Algorithm

Abstract

Examination Timetabling Problem (ETP) is a part of the Timetabling problem, which belongs to the set of combinatorial optimization problems as well as is of great significance for each University. It is addressed by heuristic and/or artificial intelligent methods. Generally, ETP involves entering a collection of exams in a specific period of time according to a set of diverse constraints that must be satisfied. Because of their wide applications, the researches are still underway on ETP. This paper presents Crow Swarm Optimization (CSO) technique which is a new intelligent way that has been proposed for the examination timetabling problems, which inspired by the behavior of crow swarms in nature. This type of crows found in North America so it is called American Crow. The CSO algorithm simulates the cooperative behavior of crows during the search for food. In nature, American Crow is divided into groups to search for food. With these strategies, it achieves the diversity in solutions and the individuals in the new algorithm explore and exploit the search space more efficiently. The research deals with the uncapacitated examination timetabling problem. The proposed algorithm is verified in 7 instances of examination timetabling. According to the results of the experiments, the proposed algorithm can provide a promising set of solutions for each examination timetabling instance by comparing our proposed algorithm with Bee Colony Optimization (BCO) and BCO with tournament selection strategy. In addition, a comprehensive comparison was conducted with the best known results on uncapacitated examination timetabling datasets, which across all the problem instances displays that the proposed approach is competitive and also works well.