S3TS: Stochastic Scenario-Structured Tree Search for Advanced Planning Under Uncertainty

2026-06-01

AI Summary

This study addresses the computational challenges arising from the coupling of nonlinear dynamics and renewable energy uncertainty in power system scheduling. To tackle this problem, the authors propose a structured scenario tree algorithm that integrates Monte Carlo Tree Search (MCTS) with stochastic optimization. This approach extends MCTS for the first time into a stochastic nonlinear programming framework, explicitly constructing a scenario tree to represent uncertainty while embedding high-fidelity nonlinear models to enable efficient decision-making. Experimental results demonstrate that under linear conditions, the proposed method achieves near-optimal operational costs with a deviation of no more than 14% from the theoretical optimum. In highly nonlinear settings, it reduces operating costs by 51% compared to myopic policies and by 5.4% relative to deterministic MCTS, substantially improving solution quality and robustness.

Abstract

Effective scheduling in the energy sector is essential to ensure the reliable operation of electrical grids and their connected assets by, for instance, optimizing the dispatch of generation units and storage systems. An effective planning strategy must (a) accommodate advanced and potentially non-linear system models -- exploiting the increasing data availability of modern grids, and (b) explicitly handle uncertainties arising, for instance, from the integration of renewable energy sources. While existing approaches can address either non-linearity (e.g., Monte Carlo Tree Search) or uncertainty (e.g., stochastic mathematical optimization), there is a lack of planning techniques capable of addressing both challenges simultaneously. To bridge this gap, we propose a Stochastic Scenario-Structured Tree Search (S3TS) algorithm that explicitly represents uncertainty through scenario trees while enabling the integration of advanced non-linear models. We evaluate S3TS on a simulated demand response signal publication problem, largely mimicking the imbalance settlement mechanism in Belgium. The results demonstrate near-optimal performance in linear, analytically tractable settings, with costs within 14% of the mathematically optimal solution conditioned to the scenario trees. In highly non-linear scenarios, S3TS significantly outperforms baseline methods, achieving cost reductions of up to 51% and 5.4% compared to a myopic algorithm and deterministic MCTS, respectively.

Problem

• uncertainty
• non-linearity
• energy scheduling
• scenario trees
• planning under uncertainty

Innovation

• Stochastic Scenario-Structured Tree Search
• scenario trees
• non-linear optimization
• uncertainty quantification
• energy scheduling