Multiagent game-theoretic robust optimization for power system planning under source-load uncertainty

基于多智能体博弈论的电力系统规划鲁棒优化方法在源负荷不确定性下的适用性

阅读:1

Abstract

The increasing penetration of variable renewable energy and the volatility of demand have amplified the importance of uncertainty-aware planning in power systems. Traditional approaches to generation and network expansion predominantly emphasize technical uncertainties associated with wind, solar, and load forecasts, while treating planning decisions as centrally coordinated. Such assumptions overlook the heterogeneous objectives and interactions of multiple stakeholders-including regulators, grid operators, renewable energy developers, and large industrial consumers-that ultimately shape the feasibility and cost-effectiveness of system expansion. This study develops a novel multi-agent game-theoretic framework for electricity system planning under source-load uncertainty, embedding stakeholder strategies into a robust optimization model. The proposed framework conceptualizes power planning as a hierarchical game, where a entity sets regulatory signals, grid operators ensure system reliability, renewable producers decide on capacity investments, and large load users respond through consumption adjustments. Their strategic interactions are modeled through a multi-layer game formulation, with each agent optimizing its own welfare function subject to operational, economic, and policy constraints. To rigorously address uncertainty, a robust optimization approach is integrated into the game, ensuring that planning outcomes remain feasible against a wide range of renewable generation variability and demand fluctuations. The robust layer captures adverse realizations of uncertainty by embedding budget-of-uncertainty sets for both renewable production and load demand, thereby producing strategies that are resilient without being excessively conservative. Case studies based on a modified IEEE benchmark system with realistic renewable and demand data demonstrate the distinct planning trajectories produced by the model. Results reveal that under robust equilibrium, coal retirements accelerate by 15-20%, while storage investments increase by 30-40% compared to nominal baselines. Load-serving entities reduce exposure to high scarcity prices by reshaping demand during peak hours, cutting tail-event prices by 20-25%. -imposed carbon penalties translate into emission reductions of 45-55% within the planning horizon, with shortfall risks limited to less than 2 GW in extreme stress scenarios. The contributions of this work are fourfold: it redefines electricity planning as a multi-agent game rather than a centralized optimization, it systematically embeds robust optimization into the strategic equilibrium, it highlights the interplay between regulatory signals and market responses, and it demonstrates how robust equilibria mitigate both physical shortfalls and economic volatility.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。