Economic dispatch with nonsmooth cost function using hybrid method

This paper presents an effective method for solving economic dispatch problem (EDP) with nonsmooth cost function using a hybrid method that integrates particle swarm optimization (PSO) with sequential quadratic programming (SQP). PSO is the main optimizer to find the optimal global region while SQP is used as a fine tuning to determine the optimal solution at the final stage. The proposed hybrid PSO-SQP method is applied to solve EDP of a test system with ten generator units. Nonsmooth cost functions of generators can occur from either having multiple fuels or experiencing vale-point effects. The cases of interest for nonsmooth cost functions in this paper are: the case where generator units have multiple fuels; and the case where generator units have both multiple fuels and valve-point effects. The results of the proposed PSO-SQP are compared with other methods to verify its effectiveness.