Abstract

Economic Load Dispatch (ELD) is a key operational task in power systems, aimed at meeting load demandat theminimum possible generation cost while strictly satisfying generator operating limits and system constraints. However, practicalELD problems are inherently non-linear and constrained, and the inclusion of transmission losses further increasestheircomplexity, making conventional optimisation techniques less effective. To address these challenges, this work presents ahybridoptimisation framework that combines Particle Swarm Optimization (PSO) with an Artificial Neural Network (ANN) forsolvingthe ELD problem of a three-generator thermal power system, explicitly considering transmission losses. In the proposed approach, the ELD problem is formulated using quadratic fuel cost functions, generator capacity limits, andtransmission losses represented through the B-coefficient method. PSO is first applied to determine high-quality optimal dispatch solutions due to its strong global search capability and reliable convergence behaviour. The optimal solutions obtainedfromPSOare then utilised to train an ANN, enabling it to learn the non-linear relationship between system load demand and correspondinggenerator power outputs. Keywords: Economic Load Dispatch; Particle Swarm Optimization; Artificial Neural Network; Hybrid PSO–ANN; TransmissionLosses; Power System Optimisation