• Boosts renewable use, cuts curtailment, and improves multi-energy reliability. • A3C multithreaded scheduler learns hydro–PV–storage complementarity under constraints. • Multi-timescale mechanism links learning scheduling and operating constraints. • Two-step correction smooths net power and cuts curtailment in high-PV periods. • Basin-company data verify stability with convergence, ablation, sensitivity tests. With the accelerating of global energy transitions and increasingly stringent climate governance, water spillage and photovoltaic (PV) curtailment in hydro–PV–storage complementary systems have become more prominent, posing a critical bottleneck to renewable energy integration and utilization. To address this challenge, this study proposes a hydro–PV–storage complementary mechanism, with its scheduling problem solved using the asynchronous advantage actor–critic (A3C) algorithm. Specifically, the scheduling problem is formulated as a Markov decision process, leveraging the multithreaded parallelism of the A3C algorithm to efficiently learn optimal scheduling policies under complex constraints. Simulations conducted in the Yarkant River Basin demonstrate that the A3C algorithm achieves rapid and stable convergence. Relative to the reference scenario, the total absorbed energy of the proposed system increases by 1.75 × 10 6 MWh, equivalent to avoiding 1.456 million metric tons of CO 2 emissions that would otherwise be produced by coal-fired power generation for the same amount of electricity. Moreover, the proposed method demonstrates strong adaptability and scalability, maintaining stable performance under various ecological flow scenarios, as well as uncertainties in reservoir inflow and PV output. As installed capacity expands, the absorbed energy increases steadily, satisfying the development requirements of large-scale clean energy bases. Overall, the results confirm that the A3C-based hydro–PV–storage coordinated scheduling approach is efficient, environmentally friendly, and practically applicable, providing methodological support for renewable energy integration, low-carbon power system operation, and the sustainable development of energy systems.
Huang et al. (Sat,) studied this question.