Received: August 8, 2017 125 International Journal of Intelligent Engineering and Systems, Vol.10, No.6, 2017 DOI: 10.22266/ijies2017.1231.14 Adaptive Weighted Improved Discrete Particle Swarm Optimization for Minimizing Load Balancing Index in Radial Distribution Network Manikandan Subramaniyan 1* Sasitharan Subramaniyan 2 Moorthy Veerasamy 3 Viswanatha Rao Jawalkar 4 1 Sathyabama University, Chennai, India 2 ABB Global Indus. & Ser. Ltd, Chennai, India 3 Swarnandhra College of Engineering & Technology, Narsapur, India 4 VNR Vignana Joythi Institute of Engineering & Technology, Hyderabad, India. * Corresponding author’s Email: svmmani79@gmail.com Abstract: In this paper a metaheuristic based newfangled adaptive weighted improved discrete particle swarm optimization (AWIDPSO) algorithm is applied to minimize the load balancing index in radial distribution network reconfiguration (RDNR) problem. It is devised as extremely nonlinear and multimodal optimization problem under practical constraints. In order to improve the solution quality the constraint violations are augmented with objective function. Further, adaptively varying inertia weight increases the possible solution in the global search space and the proposed algorithm has obtained the optimal solution within lesser executing time. In this study, 33-bus system is analyzed for optimal network reconfiguration using the developed framework. Comparison of the simulated results with the results of well known prudent optimization technique confirms the applicability of the AWIDPSO algorithm for RDNR problem. Keywords: Metaheuristic algorithm, Radial distribution system, Network reconfiguration, Load balancing index, Voltage profile improvement. 1. Introduction In the rapidly growing research field emerging optimization technique facilitates to find a feasible solution for complex engineering problems. In which metaheuristic optimization is one that deals with optimization problems using metaheuristic algorithms. These are the simplest sense, gradient- free, non-deterministic, not problem specific and have been inspired by the natural selection process. Moreover, randomness features, intensification, and diversification driving forces of the metaheuristic algorithms bring the control parameters of the nonlinear problem to the edge, whereas, mathematical methods difficult to produce an accurate result. So, the metaheuristic optimization is to be an effective tool to solve nonlinear problems [1]. As stated above it is not a problem specific, has less hold on initial solution point and tendency to solve large-scale and any kind of complex engineering problem, the researchers are motivated to solve distribution network reconfiguration (DNR) problem. It can be mathematically formulated as an optimization problem subjected to various operational constraints to ascertain radiality in the distribution network optimally that reduces the power loss and minimizes load balancing index [2- 3] or maximize the benefits under the normal operation conditions [4]. In this context some metaheuristic algorithms such as genetic algorithm (GA) [5], a new hybrid evolutionary algorithm (EA) based on the combination of the honey bee mating optimization (HBMO) and the particle swarm optimization (PSO) [6], plant growth simulation algorithm [7], modified