Covariance matrix adaptation evolution strategy based on multidistribution collaborative sampling

Abstract:
In the power system, multiple heuristic optimization algorithms optimize the economic dispatch (ED) of multiple fuel options, but they all have the disadvantage of low convergence accuracy. This paper proposes an covariance matrix adaptation evolution strategy based on multi-distribution collaborative sampling (MD-CMA-ES) to solve the large-scale ED problem. Multi-distribution cooperative sampling uses the accuracy of normal distribution and the breadth sampling of uniform distribution to improve sample diversity and prevent falling into the local optimum. The method is validated on 2 test systems consisting of 10 and 40 generator sets and compared with other algorithms, and the experimental results show that the method can obtain better solutions.