Comparative performance evaluation of economic load dispatch using metaheuristic techniques: A practical case study for Bangladesh

The Economic Load Dispatch (ELD) problem is a critical optimization challenge in power systems, aiming to minimize operational costs while meeting power demand. Traditional methods, such as the Gradient and Lambda-iteration techniques, often fail to address the non-linearities and complexities of modern power systems. This study evaluates three metaheuristic algorithms—Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Harmony Search (HS)—for solving the ELD problem in a thermal power plant in Bangladesh. The algorithms are tested on systems with 6 generating units (600 MW demand) and 15 generating units (1275 MW demand), incorporating constraints such as prohibited operating zones, ramp rate limits, and transmission losses. The results show that PSO achieves the lowest power losses (0.71 MW for 6 units and 1.5 MW for 15 units) and the lowest generation costs ($8989.24/h for 6 units and $10,550.50/h for 15 units). HS demonstrates the fastest convergence, with computational times of 3.9 s (6 units) and 8.0 s (15 units), but incurs higher generation costs ($8994.18/h for 6 units and $11,050.80/h for 15 units) and power losses (8.31 MW for 6 units and 6.0 MW for 15 units). ACO provides a balanced approach, with moderate computational times (6.5 s for 6 units and 12.5 s for 15 units) and competitive generation costs ($8989.40/h for 6 units and $11,000.32/h for 15 units). This study highlights the trade-offs between computational efficiency and solution quality, offering practical insights for power system operators. PSO emerges as the most effective algorithm for minimizing costs and losses, while HS is suitable for time-sensitive applications. These findings provide a foundation for future research into hybrid optimization techniques, advancing the efficiency and cost-effectiveness of power generation in Bangladesh and beyond.