Evolutionary Learning Based Optimization for Solving Non-convex Optimal Power Flow Problems: A Modern Teaching–Learning-Based Optimization Algorithm
Year: 2026
Type: preprint
Abstract: Optimal power flow (OPF) is one of the fundamental and complex problems in power system planning and operation, which requires efficient and flexible optimization methods due to its nonlinear characteristics, multiple constraints, and multi-objective nature. In this paper, a new meta-heuristic algorithm named evolutionary learning based optimization (ELBO) is presented, which is an extended and improved version of the teaching learning based optimization (TLBO) algorithm. In the design of ELBO, an evolutionary step with a trigonometric mutation mechanism is added to the TLBO training process to reduce premature convergence and maintain population diversity in the search space. The proposed algorithm has been evaluated in two main areas: first, a set of CEC2014 benchmark test functions and several real engineering optimization problems were first used. In these experiments, ELBO was able to show superior performance over modified versions of TLBO and other reference algorithms in accuracy, convergence speed, and stability. Second, in solving several OPF optimization problems on standard 30- and 57-bus test networks, in all of which the ELBO algorithm has been able to provide more accurate, more stable, and faster convergence rates than other methods. These results confirm the effectiveness and high power of ELBO in solving complex and real problems in the fields of electrical engineering and optimization, and introduce it as an innovative and powerful tool.
Source: Research Square
Authors: Roqia Rateb, Maytham N Meqdad, Taybeh Salehnia
Institutions: Hashemite University, Al-Mustaqbal University, Razi University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7203780656