arXiv subjects

Luca Santosuosso

Publications and source records attributed to Luca Santosuosso.

8 recordsLinked to original sources

arXiv2026

Stochastic Virtual Power Plant Dispatch via Temporally Aggregated Distributed Predictive Control with Performance Guarantees

This paper addresses the energy dispatch of a virtual power plant comprising renewable generation, energy storage, and thermal units under uncertainty in renewable output, energy prices, and energy demand. The nonlinear dynamics and multiple sources of uncertainty render traditional stochastic model predictive control (MPC) computationally intractable as the dispatch horizon, scenario set, and asset portfolio expand. To overcome this limitation, we propose a novel controller that seamlessly integrates MPC with time series aggregation and distributed optimization, simultaneously reducing the temporal, asset, and scenario dimensions of the problem. The resulting controller provides a rigorous performance guarantee through theoretically validated bounds on its approximation error, while leveraging dual information from previous MPC iterations to adaptively optimize the temporal aggregation. Numerical results show that the proposed controller reduces runtime by over 50% relative to traditional stochastic MPC and, crucially, restores tractability where the full-scale dispatch model proves intractable.

math.OC↗

arXiv2026

Unlocking the Informational Value of Marginal Costs for Exact Time Series Aggregation in Generation Expansion Planning

This paper addresses the generation expansion planning (GEP) problem, formulated as a mixed-integer linear programming model with intertemporal storage constraints. Being generally NP-hard, the problem's computational complexity grows sharply with the planning horizon and the number of binary variables. While previous research has tackled this challenge using heuristic time series aggregation (TSA) methods, we propose a theoretically grounded marginal-cost-based TSA, designed to construct an aggregated model that preserves the active constraints of its full-scale counterpart, thereby explicitly targeting exact temporal aggregation. This TSA method is embedded within solution algorithms that iteratively refine theoretically validated bounds on the maximum error introduced by the temporal aggregation relative to conventional full-scale optimization, thus offering a formal performance guarantee to the decision-maker. Numerical results highlight the computational advantages of the proposed algorithms, which notably recover tractability whereas full-scale optimization proves intractable.

math.OC↗

arXiv2026

Distributed Stochastic Model Predictive Control with Temporal Aggregation for the Joint Dispatch of Cascaded Hydropower and Renewables

This paper addresses the real-time energy dispatch of a hybrid system comprising cascaded run-of-the-river hydropower plants, wind, and solar photovoltaic units, operated under uncertainty in water inflows and renewable power generation. Traditional scenario-based stochastic model predictive control (MPC) schemes suffer from severe computational limitations due to the high dimensionality induced by both the temporal and scenario dimensions of the dispatch problem, as well as the inherent nonconvexities associated with cascaded hydropower dynamics. To overcome these challenges, we propose a novel control scheme that seamlessly integrates time series aggregation (TSA), distributed optimization, and stochastic MPC. The resulting temporally aggregated distributed stochastic MPC scheme simultaneously reduces the temporal dimension of the dispatch problem via TSA and decomposes it across scenarios through distributed optimization. Our main theoretical result establishes a formal performance guarantee for the proposed controller, enabling a rigorous quantification of its solution accuracy at every MPC iteration. Numerical results based on a real-world case study show the effectiveness of the proposed controller, achieving up to 74% reduction in computational effort relative to the full-scale centralized counterpart when the required solution accuracy is at least 99%, and up to 85% when the accuracy requirement is relaxed to 95%. Notably, the proposed controller not only significantly enhances computational efficiency relative to the traditional full-scale centralized counterpart, but more importantly restores computational tractability, whereas the traditional controller fails to solve the dispatch problem within the prescribed time limit for computing control actions.

math.OC↗

arXiv2026

Machine Learning for Exact Time Series Aggregation in Generation Expansion Planning with Energy Storage

This paper investigates a generation expansion planning (GEP) problem encompassing renewable, thermal, and storage technologies while simultaneously optimizing market participation, operational expenditures, and capital investment. To alleviate the computational burden of the GEP model, we propose a novel iterative time series aggregation (TSA) method that constructs a temporally aggregated counterpart of the original full-scale GEP model. Unlike traditional TSA methods, which are purely heuristic, our method enables the assessment of the optimality gap between the aggregated and full-scale models. Moreover, by leveraging machine learning-based estimates of the GEP model marginal costs, the algorithm guides TSA to construct an aggregated model that preserves the active constraints of its full-scale counterpart, which has been shown to yield exact temporal aggregation. Numerical results show that incorporating estimated marginal costs as clustering features substantially improves the quality of temporal aggregation compared with traditional TSA methods that rely solely on input data analysis.

math.OC↗

arXiv2026

Towards Exact Temporal Aggregation of Time-Coupled Energy Storage Models via Active Constraint Set Identification and Machine Learning

Time series aggregation (TSA) aims to construct temporally aggregated optimization models that accurately represent the output space of their full-scale counterparts while using a significantly reduced temporal dimensionality. This paper presents a theoretical approach that achieves exact temporal aggregation of full-scale power system models -- even in the presence of energy storage time-coupling constraints -- by leveraging active constraint sets and dual information. This advances the state of the art beyond existing TSA methods, which typically cannot guarantee solution accuracy or rely on iterative procedures to determine the required number of representative periods. To bridge the gap between this theoretical analysis and practical application, we employ machine learning, i.e., classification and clustering, to inform TSA in models that co-schedule variable renewable energy sources and energy storage. Numerical results show substantially improved computational performance relative to the full-scale model, while maintaining a favorable trade-off between solution accuracy and complexity.

math.OC↗

arXiv2025

A Scenario-Spatial Decomposition Approach With a Performance Guarantee for the Combined Bidding of Cascaded Hydropower and Renewables

This study develops a scalable co-optimization strategy for the joint bidding of cascaded hydropower, wind, and solar energy units, treated as a unified entity in the day-ahead market. Although hydropower flexibility can manage the stochasticity of renewable energy, the underlying bidding problem is complex due to intricate coupling constraints and nonlinear dynamics. A decomposition in both scenario and spatial dimensions is proposed, enabling the use of distributed optimization. The proposed distributed algorithm is eventually a heuristic due to non-convexities arising from the system's physical dynamics. To ensure a performance guarantee, trustworthy upper and lower bounds on the global optimum are derived, and a mathematical proof is provided to demonstrate their existence and validity. This approach reduces the average runtime by up to 35% compared to alternative distributed methods and by 57% compared to the centralized optimization. Moreover, it consistently delivers solutions, whereas both centralized and alternative distributed approaches fail as the size of the optimization problem grows.

math.OC↗

arXiv2025

What Are We Clustering For? Establishing Performance Guarantees for Time Series Aggregation in Generation Expansion Planning

Generation expansion planning (GEP) is a prominent example of capacity expansion problems in operations research. Being generally NP-hard, GEP optimization models can become intractable when nonconvex dynamics, time-coupling constraints, and complex asset interactions are involved. Time series aggregation (TSA) tackles this by reducing temporal complexity via input data clustering. However, existing TSA methods either focus solely on preserving the statistical features of the input data, yielding heuristics without guarantees on the aggregated model's accuracy, or provide error bounds limited to linear models, neglecting time-coupling constraints and applying only to specific clustering techniques. Moreover, these bounds typically pertain solely to the GEP objective function and do not extend to other stakeholder-specific metrics, such as decision vector partitions. To tackle these issues, we demonstrate that an appropriately constructed aggregated model always provides a lower bound on the optimal objective function value of the full-scale GEP model in both mixed-integer linear and mixed-integer quadratic formulations with time-coupling, independent of the clustering technique employed. Building on this, we propose a performance-guaranteed TSA-based solution algorithm that iteratively refines objective function bounds while generating feasible solutions to the full-scale model at each iteration. We then discuss a comparison with Benders decomposition and demonstrate how the derived bounds can be extended to error estimates for stakeholder-specific metrics. Numerical results show the computational advantages of our method over both full-scale optimization and classical Benders decomposition.

math.OC↗

arXiv2025

Optimal Virtual Power Plant Investment Planning via Time Series Aggregation with Bounded Error

This study addresses the investment planning problem of a virtual power plant (VPP), formulated as a mixed-integer linear programming (MILP) model. As the number of binary variables increases and the investment time horizon extends, the problem can become computationally intractable. To mitigate this issue, time series aggregation (TSA) methods are commonly employed. However, since TSA typically results in a loss of accuracy, it is standard practice to derive bounds to control the associated error. Existing methods validate these bounds only in the linear case, and when applied to MILP models, they often yield heuristics that may even produce infeasible solutions. To bridge this gap, we propose an iterative TSA method for solving the VPP investment planning problem formulated as a MILP model, while ensuring a bounded error in the objective function. Our main theoretical contribution is to formally demonstrate that the derived bounds remain valid at each iteration. Notably, the proposed method consistently guarantees feasible solutions throughout the iterative process. Numerical results show that the proposed TSA method achieves superior computational efficiency compared to standard full-scale optimization.

math.OC↗