Comparative Application of Multi-Objective Evolutionary Algorithms to the Voltage and Reactive Power optimization Problem in Power Systems

Abstract

This study investigates the applicability of two elitist multi-objective evolutionary algorithms (MOEAs), 
namely the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) and an improved Strength Pareto 
Evolutionary Algorithm (SPEA2), in the voltage and reactive power optimization problem. The problem 
has been formulated mathematically as a nonlinear constrained multi-objective optimization problem 
where the real power loss, the load bus voltage deviations and the installation cost of additional reactive 
power (VAR) sources are to be minimized simultaneously. To assess the effectiveness of the proposed 
approach, different combinations of the objectives have been minimized simultaneously. The simulation 
results showed that the two algorithms were able to generate a whole set of well distributed Pareto-optimal 
solutions in a single run. Moreover, fuzzy logic theory is employed to extract the best compromise solution 
over the trade-off curves obtained. Furthermore, a performance analysis showed that SPEA2 found better 
convergence and spread of solutions than NSGA-II. However, NSGA-II found more extended trade-off 
curves in some cases and required less computational time than SPEA2.