Comparison of Algorithms to Solve Multi-objective Optimal Reactive Power Dispatch Problems in Power Systems with Nonlinear Models and a Mixture of Discrete and Continuous Variables

Abstract

Optimal reactive power dispatch (ORPD) is a way to improve power system performance. Determination of the optimal value of the control variable can reduce the objective function to be achieved. The optimization of the two objective functions simultaneously is called multi-objective ORPD (MORPD). This research has a major contribution in proposing and utilizing original ideas from new algorithms and new ideas from the old algorithm to solve more complex variables and challenging ORPD problems. The ORPD problem is formulated as a nonlinear model with variables consisting of continuous and discrete. The proposed multi-objective algorithm is time-varying particle optimization (MOTVPSO), ant lion objective (MOALO), dragonfly algorithm (MODA), grey wolf optimizer (MOGWO), and multi-objective multi-verse optimization (MOMVO). To measure the effectiveness of those algorithms, testing is performed on the IEEE 57-bus. The simulation results show that the MOTVPSO algorithm can contribute more dominantly from the statistical tests conducted compared to previous studies and all four algorithms in this work to minimize real power loss. Whereas the MOMVO has an advantage in computational time efficiency.