Genetic algorithm solution to economic dispatch problems

This paper presents an economic dispatch algorithm based on the generic alg orithm (GA) for the determination of rite global or quasi-global optimum di spatch solution under consideration of transmission losses. In practice, th e efficiency of GA is sometimes hindered by a poor performance in a localiz ed search or by the difficultly of finding and maintaining feasibility for a constrained problem. To deal with these problems, an advanced genetic alg orithm is developed. With the help of local improvement operator; combined with a conjugate gradient method, the performance of the GA is enhanced. By means of heuristic scaling of fitness functions and adaptation of penalty factors, the genetic search in the feasible region can be improved. The alg orithm is implementation in the binary search space and tested with multi-m odal non-linear functions. The application of the algorithm to a test syste m is also demonstrated, where transmission losses are represented by the B- matrix loss formula.