Wind power fitness function calculation based on niche genetic algorithm

Abstract

According to the function of wind speed and wind power made by least square regression, we use genetic algorithm (GA) for one-dimensional test. The meanwhile, based on the impact of wind speed and temperature relation in two-dimensional function, we apply niche genetic algorithm (NGA) for training. We use eight combinations: elitist, adaptive crossover and mutation probability; niche algorithm; pre-selection algorithm; penalty function algorithm; niche algorithm and pre-selection algorithm combination; niche algorithm and penalty function; pre-selection algorithm and penalty function algorithm; niche algorithm, pre-selection algorithm and penalty function algorithm for joint use. In comparison, the niche algorithm, pre-selection algorithm, genetic algorithm penalty function used in conjunction iteration is better. The average fitness of two-dimensional niche genetic algorithm iteration is smooth and basically stable. One-dimensional NGA fitness situation: fval =85.4512. The elapsed time is 8.073602s. The final fitness value of two-dimensional fitness function is 749.9563. (5 pages)