Surrogate-assisted Optimization Method for Optimal Dynamic Economic Dispatch including Aggregation Regulation of Wind and Biomass Farms

Due to the uncertain nature of wind and the booming number of generator units, a novel dynamic economic dispatch (DED) approach involving wind-biomass aggregation regulation strategy is proposed. To solve this DED problem, a novel surrogate-based optimization algorithm is proposed. On one hand, an innovative chaotic map (named exponent-derivativeremainder (CDR) chaotic map) is proposed and integrated into Young's double-slit experiment (YDSE) optimizer to enhance the convergence of algorithm and obtain optimal dispatching decision. On the other hand, a multi-head attention neural network (MANN) surrogate model is proposed to approximate original objective function in DED problem for reducing the evaluation time of fitness functions and total running time. Furthermore, a novel surrogate model management scheme is proposed to reduce the error of fitting MANN surrogate model. The effectiveness of proposed method is verified by a forty-unit system with wind farms and biomass/bagasse farms.