Improved Genetic Algorithm for Substation Location Optimization

Aiming at the substation location optimization with multi constraints and multi indexes, an improved genetic algorithm is proposed. Combined with the characteristics and requirements of substation location, the real number coding strategy and elite retention scheme are adopted. The minimum annual cost of substation planning is regarded as the fitness to realize the adaptive search of spatial solution in spatial range, which can effectively solve the problems of local optimal solution and premature. The probability of mutation and crossover is determined according to individual fitness differentiation, which improves the convergence speed and solution accuracy of the algorithm. The example results show that the algorithm has good optimization ability, convergence characteristics, simple operation and fast running speed. It can better meet the needs of substation location selection.