Application of particle swarm optimization technique to a profit-based unit commitment problem

This paper aims at investigating the performance of a Particle Swarm Optimization (PSO) algorithm combined with Lagrange Relaxation (LR) method, called LR-PSO, so as to solve a profit-based Unit Commitment (UC) problem. For the proposed LR-PSO method, PSO is applied to update the Lagrange multipliers and is also incorporated into the LR method to improve its performance. In addition, the proposed approach also utilized the Gradient method to adjust another Lagrange multiplier. That will lead to increase in the searching potential. It can be concluded from the comparison of simulation results between the profit-based UC with and without Gradient method that the proposed LR-PSO method with Gradient method can achieve the optimum solution, while improving the convergence capability.