Short-term economic environmental dispatch of hydrothermal power system based on predictive search NSGA-II (iSPEC 2020)

Abstract

NSGA-II uses genetic search and non-dominated sorting strategies to find the Pareto front. The search direction is not active, which affects the efficiency of searching for the optimal Pareto front. This paper first adopts uniformly distributed sample initialization for the selection of the initial population to ensure the uniformity of the initial solution; secondly, in the process of constructing the offspring population, adopts the elite guidance strategy to accelerate the convergence of the algorithm; finally, according to the distribution characteristics of each generation of population samples, The solution set with Pareto grades 1 and 2 predicts the non-dominated direction of the Pareto front, and actively searches for a step toward the non-dominated direction from each sample of the Pareto front, finds better possible non-dominated solutions and participates in the next generation subgroup reconstruction. The improved NSGA-II algorithm of direction prediction search is applied to the short-term economic and environmental dispatching of hydrothermal power system. The simulation results verify the feasibility and effectiveness of the algorithm and the constraint processing method, and provide an efficient new method for solving the multi-objective scheduling problem of hydrothermal power systems.