Adaptive quantum-inspired evolutionary algorithm for optimizing power losses by dynamic load allocation on distributed generators

Abstract

In this paper, an Adaptive Quantum-inspired Evolutionary Algorithm (AQiEA) \n has been applied for minimizing the power losses in the distribution network \n by suitable placement, sizing and subsequent allocation of load on \n Distributed Generators (DG) for a varying load with a time horizon of \n twentyfour hours. Many efforts have been reported in the literature to \n minimize power losses. However, they have mostly used a fixed load, i.e., \n nonvarying load, whereas it is well known that load in distribution network \n varies during the day. An investigation was undertaken to find the reduction \n in power losses on a timevarying load. It has been found that the average \n power losses for dynamic load allocation on DGs for every hour have a \n maximum reduction in power loss as compared with other well-known cases in \n the literature. Optimal location and size of DG is a difficult nonlinear, \n non-differentiable combinatorial optimization problem. AQiEA is used to find \n the appropriate location and capacity of DG for a varying load with a time \n horizon of twenty-four hours to minimize the power losses. AQiEA doesn’t \n require additional operators like local search and mutation to improve the \n convergence rate and avoid the premature convergence. A Quantum Rotation \n inspired Adaptive Crossover operator is used as a variation operator, which \n is parameter free. The effectiveness of AQiEA is demonstrated on two test \n bus systems viz., 33 bus system and 69 bus system, which are used as \n benchmark problems for validating the proposed methodology as well as for \n comparative testing amongst existing techniques. Wilcoxon signed rank test \n is also used to demonstrate the effectiveness of AQiEA. The experimental \n results show that AQiEA has better performance as compared to some existing \n ‘state of art’ techniques.