Optimal generation expansion planning via improved genetic algorithm approach

This paper presents an improved genetic algorithm approach developed to solve the optimal generation expansion planning problem of an all-thermal power system. The problem is focused on the optimal mix of generation units in a given target year with the constrained consideration of certain thermal units committed during peaking periods. The problem formulation thus requires considering the technical limits of the thermal unit outputs due to the large difference between the daily peak-load and valley-load demands. In addition, the implementation issues of penalty coefficients, ranking, adaptive crossover and mutation probabilities are effectively considered in the algorithm. The test results on a 14-generator power system are presented. The results show that the methodology is effective in solving such mixed integer, constrained nonlinear generation expansion problem.