Given the high coverage and zoning of new-energy, the auxiliary-service dispatching in the southern power market faces many constraints and a large-scale problem. Construct a three-layer model of supply and demand – service in this paper and reduce it to a determination problem, thus demonstrating that the problem is NP-hard. Based on the above, a hybrid algorithm of genetics and particle swarm is proposed. Rule initialization and repair guarantee feasibility, and a penalty function is employed for approximation to balance the global and local optima. According to the simulation set in the south, multi-scale experiments were carried out. Based on the above experiments, the total cost for the medium-sized scenario was 12,435 yuan; the constraint satisfaction rate reached 100%, and the number of convergence rounds was as low as 132, which outperformed GA, PSO, and ACO. Innovation is to: provide proof of complexity; propose a reproducible hybrid framework; strike a balance between efficiency, economy and fairness; maintain minute-level efficiency; be consistent across scales; have excellent robustness and be deployable.
