Random walk with variable step size as mutation operator of genetic algorithm for solving combined heat and power economic dispatch problem

In this paper, a mutation operator based on the random walk with variable step size utilized by genetic algorithm (GA) is under study. Proposed method is used to solve combined heat and power economic dispatch (CHPED) problem. Combined heat and power economic dispatch problem is a complex and complicated optimization problem, which has non-convex, nonlinear and non-smooth objective function and many equality and inequality constraints. Normally, the step size of the random walk is obtained by probability distributions including exponential family and heavy-tailed distributions, which are both used in implementations of this paper. The proposed method achieved from different distributions, as determiner of step size in random walk based mutation, are implemented on the problem. The results are compared with numerous heuristic algorithms available in the literature. Also, convergence characteristics obtained from different mutation operators are compared. The analysis of achieved results and their improvements shows the capability of the random walk as mutation operator of GA.