Application of a Novel Jaya Algorithm Based on Chaotic Sequence and Opposition-based Learning in the Multi-objective Optimal Operation of Cascade Hydropower Stations System

Abstract

The traditional operation of the cascade hydropower stations system (CHPS) mainly focus on the maximization of power generation benefits, but ignores the interference of CHPS operation to the river ecosystem, therefore, carrying out the multi-objective optimal operation (MOOP) of CHPS considering ecological demands is crucial. In this paper, a MOOP model considering the ecological objective is established. To effectively solve the MOOP problems, a novel multi-objective Jaya algorithm (MOCOM-Jaya) is proposed, where the quality of the initial population is enhanced based on the chaotic sequence, the later disturbance term and Gaussian mutation are incorporated to improve the local search ability, the elite opposition-based learning is adopted to broaden the optimization space. The proposed algorithm is applied to the study of MOOP of CHPS in the Wujiang river, and the results show that compared with MOPSO and NSGA-II, MOCOM-Jaya can gain the solution set with better convergence and distribution for the MOOP. The competition relationship between the power generation objective (PGO) and the ecological objective (ECO) is revealed based on the partial replacement ratio method. The results show that the competitiveness of PGO and ECO experienced a trade lead with the increase of power generation. The mean competitiveness ratios of PGO to ECO ( $\overline {CP{R_{P - E}}} $ ) in three typical years (dry, normal, wet) are 3.22, 3.17 and 3.15, indicating that the PGO is dominant in the competition with the ECO as a whole.