EFFECTS OF EQUALITY CONSTRAINT HANDLING ON ECONOMIC DISPATCH USING DIFFERENTIAL EVOLUTION ALGORITHMS

摘要

Heuristic optimization is an efficient method to solve non-convex economic dispatch (ED) problems. In the past decade, several heuristic algorithms have been developed to improve quality of the ED result. However, those improvements are still inconclusive whether they are contributed to the power of the optimization algorithm itself or the influence of the constraint handling technique. In this paper, the hypothesis is that the strategy for dealing with the power balance (equality) constraint has the greater influence to the solution quality than the optimization algorithm. Therefore, this paper investigates the effects of this strategy on reliability and accuracy of the ED result. Four different DE algorithms are applied as the optimizers. The simulation results on a 15-generator power system confirm that the formulated hypothesis is true. It is also found out that with such a strategy a greedy algorithm searching around the local best converges faster than other algorithms.