An enhanced micro-PSO method to deal with asymmetric electricity markets competition within hydropower cascade

Electricity deregulation has intensified competition in the clean energy sectors, notably within cascaded hydropower systems. Due to the natural interdependencies among hydropower stations, upstream stations with self-interested motivations can significantly influence downstream operations, creating asymmetrical competition and considerable uncertainty for downstream stations’ bidding in the electricity market. To address this issue, this paper, from the perspective of the downstream stations, presents a bilevel model to estimate the market strategies of upstream competitors using inverse optimization and non-private historical data, with a detailed consideration of their hydraulic connections. In the bilevel structure, the upper-level model aims to estimate key parameters of the rivals’ generation function, while the lower-level model simulates their bidding behavior using parameters provided by the upper-level model. Engineering experience is incorporated to streamline the decision variables of the upper-level model into only four. To solve this non-convex model, an Enhanced Micro Particle Swarm Optimization (EMPSO) algorithm is proposed, which employs a tabu table-based reinitialization strategy to ensure diversity within the small population and introduces mutation operations to enhance exploration capability. Applications in the Lancang River basin in China demonstrate the model’s precision, efficiency, and stability. Specifically, the estimated errors for the rival’s generation and power discharge are all within 1%, while the estimated errors for all of the four decision variables are all within 2%. Additionally, it is found that a strong correlation between the objective function and parameter values. The algorithm maintains robust search capability throughout the entire process. Finally, statistical tests confirm the significant superiority of EMPSO over conventional methods.