Comparative Analysis of Nature-Inspired Algorithms for Optimal Power Flow Problem: A Focus on Penalty-Vanishing Terms and Algorithm Performance

Abstract

This study presents a comparative analysis of multiple nature-inspired algorithms for solving the non-polynomial Optimal Power Flow (OPF) problem. Through numerical evaluations, we assess their performance across diverse objective functions, addressing complexities such as multi-fuel sources, valve point effects, and prohibited zones. The study involves the implementation of different nature-inspired heuristics and variants of the differential evolution algorithm to analyze their efficacy in solving the OPF problem within the context of large networks, specifically IEEE-30 and IEEE-57. The objectives of this research are threefold: (i) to determine the most effective nature-inspired algorithms for each case under consistent constraints, initial conditions, and using optimized parameters, (ii) to assess the success rate of penalty-vanishing terms concerning the penalized function versus the actual objective function, and (iii) to explore the impact of minor variations within a network on the behaviors, results, and profiles of penalty-vanishing terms. Utilizing a low-high sorting ranking method, considering mean, maximum, and minimum values for result computation and sorting, we identify the optimal algorithm among all those assessed for various objective functions, alongside assessing the success rate of penalty-vanishing terms. Our findings reveal that the differential evolution algorithm best version (DEAB) emerges as the most valuable solution.

Keywords

• Nature-inspired algorithms
• optimal power flow
• optimization
• penalty-vanishing terms
• success rate