This paper proposes a two-stage Stackelberg game model based on the combination of adaptive weighting factors and penalty function relaxation, in which the power plant is regarded as the dominant party for capacity investment and price setting, and the load-side consumers are regarded as the passive party for purchasing electricity, which constitutes a two-stage optimization framework that characterizes the behavioral characteristics of power producers at the upper level, and the response characteristics of consumers’ demand at the lower level. This constitutes a two-stage optimization framework in which the upper model characterizes the behavior of generators and the lower model characterizes the demand response characteristics of consumers. To address the high complexity of the above model, this paper proposes a hybrid genetic-particle swarm decomposition coordinated intelligent optimization algorithm based on the KKT optimality condition, which utilizes the global optimization ability of the genetic algorithm and the strong local optimization ability of the particle swarm algorithm for coordinated optimization, and transforms the original problem into a single-layer mixed-integer planning model. At the same time, parallel computing is utilized to greatly improve the solving efficiency and convergence speed of the algorithm. In summary, under different types of market models, the modified model proposed in this paper has good robustness and can improve the accuracy and fairness of capacity cost recovery, thus enhancing the satisfaction of all market players and the economy of the whole system.
