Differential evolution based on migrating variables for the combined heat and power dynamic economic dispatch

Abstract

A differential evolution using migrated variables is proposed to deal with the combined heat and power dynamic economic dispatch problems in this paper. The new differential evolution improves the classical one from two aspects. Firstly, it incorporates an attracting factor involving direction information into the mutation operation, providing mutant vectors with more opportunities of searching potential regions. Secondly, it replaces a number of segments of the mutant vectors with those of previous target ones, contributing to improvements of candidates. Additionally, a method of repairing solutions is proposed to assist the solutions in moving towards the feasible regions rapidly. Each repaired solution can always satisfy four kinds of constraints including power generation limits, capacity limits of combined heat and power units, heat generation limits and ramp rate limits. In addition, it is likely to satisfy the other three kinds of constraints including prohibited operating zones, power balances and heat balances. As a result, the new differential evolution combined with the method of repairing solutions is able to accelerate the eliminations of constraint violations and the reduction of objective function value for each solution. Experimental results demonstrate that the new differential evolution can obtain desirable results and it outperforms the other five algorithms for the eight combined heat and power dynamic economic dispatch cases with different dimensions.