A Learning based Evolutionary Approach for Minimization of Matrix Bandwidth Problem

Nowadays, graphs and matrix have been used extensively in computing. In this paper an evolutionary approach to solve a problem related with the matrix called minimization of bandwidth problem is proposed. Due to difficulties to solve of this problem, using of evolutionary processing and especially genetic algorithm is efficient. In this paper by adding learning concepts such as penalties and rewards (Guidance) to the genetic algorithm, we obtained an efficient method for solving minimization of matrix bandwidth problem; so that in search process, the speed of finding the answer to the significantly increased. Obtained results of experiments on twenty sample matrices show the efficiency and speed of suggested method in comparison with other methods.