Comparative analysis of enhanced PSO and MILP methods for cost-effective transmission congestion management in deregulated power systems

Abstract

This study addresses transmission congestion management in deregulated power systems through generator rescheduling, aiming to minimize active and reactive adjustment costs while maintaining reliable operation. Two optimization approaches, an enhanced Particle Swarm Optimization (PSO) algorithm with adaptive parameter tuning to improve convergence stability, and a Mixed-Integer Linear Programming (MILP) formulation based on DC-OPF linearization, are developed and compared. The methods are tested on IEEE 14-, 30-, and 118-bus systems. Both approaches effectively balance system demand, but MILP consistently achieves lower rescheduling costs, reduced network losses, and improved voltage profiles across the three test cases. PSO incurred active losses of 4.7%, 11.03%, and 10.87%, and reactive losses of 3.67%, 15.39%, and 12.31%, whereas MILP achieved active losses of 5%, 15.5%, and 12.5%, and reactive losses of 5%, 24%, and 13%, demonstrating superior cost efficiency and scalability. Statistical analysis using the Wilcoxon signed-rank test further confirms MILP’s scalability advantage, while the enhanced PSO demonstrates competitive performance under nonlinear conditions. Overall, MILP provides a more cost-effective and robust solution, with future work extending the framework to probabilistic models for renewable integration.