A genetic algorithm by using a matrix crossover for determining classroom seating arrangements

Classroom seating arrangements are important topics that affect each and every students and teachers. In this study we propose a method for determining the classroom seating arrangements considering relationships between students. In order to determine the optimal classroom seating arrangements, a genetic algorithm is applied on the basis of the questionnaire results and the analysis of the observation of behaviors between students. In the genetic algorithm, we propose a procedure named a matrix crossover. Our optimization problem is to determine the combination of classroom seating arrangements in such a way that the minimum of fitness rates of all students is as large as possible. Based on our proposed classroom seating arrangements, we carried out experiments at a national college of technology in Japan. It is found from the experiments that satisfaction of each student for our proposed method is higher than the traditional one. It is also found from another questionnaire that each student is actually comfortable in the classroom because of the relations between other students. It is confirmed from the experimental results that the proposed method was effective.