arXiv subjects

Pietro Favaro

Publications and source records attributed to Pietro Favaro.

3 recordsLinked to original sources

arXiv2025

Accelerating Underground Pumped Hydro Energy Storage Scheduling with Decision-Focused Learning

Underground pumped hydro energy storage (UPHES) systems play a critical role in grid-scale energy storage for renewable integration, yet optimal day-ahead scheduling remains computationally prohibitive due to nonlinear turbine performance characteristics and discrete operational modes. This paper presents a decision-focused learning (DFL) framework that addresses the computational-accuracy trade-off in UPHES day-ahead scheduling. The proposed methodology employs neural networks to predict penalty weights that guide recursive linearization, transforming the intractable MINLP into a sequence of convex quadratic programs trained end-to-end via differentiable optimization layers. Case studies across 19 representative Belgian electricity market scenarios demonstrate that the DFL framework effectively navigates the trade-off between solution quality and computation time. As a refiner, DFL adds approximately 1.2 s of post-processing to improve profit by 1.1% over piecewise MIQP baselines. As a real-time scheduler initialized with a fast global linear MIQP, the end-to-end time is 3.87 s (a 300-fold speedup) while maintaining profitability within 3.6% of the piecewise MIQP benchmark. An MIP-free deployment mode, which warm-starts from a historical schedule database and thus needs no mixed-integer solve, retains 92% of benchmark profit on held-out price scenarios.

eess.SY↗

arXiv2025

Decision-Focused Learning for Neural Network-Constrained HVAC Scheduling

Heating, Ventilation, and Air Conditioning (HVAC) is a major electricity end-use with a substantial potential for providing grid services, such as demand response. Harnessing this flexibility requires accurate modeling of the thermal dynamics of buildings, a difficult task because nonlinear heat transfer and recurring daily cycles make historical data highly correlated and insufficient to generalize to new weather, occupancy, and control scenarios. This paper presents an HVAC management system formulated as a Mixed Integer Quadratic Program (MIQP), where Neural Network (NN) models of thermal dynamics are embedded as exact mixed-integer linear constraints. Unlike traditional training approaches that minimize prediction errors, we employ Decision-Focused Learning (DFL) to learn the NN parameters with the objective of directly improving the HVAC cost performance. However, the discrete nature of MIQP hinders DFL, as it leads to undefined and discontinuous gradients, thus impeding standard gradient-based training. We leverage Stochastic Smoothing (SS) to enable efficient gradient computation without the need to differentiate the MIQP. Experiments on a realistic five-zone building using a high-fidelity simulator demonstrate that the proposed SS-DFL approach outperforms conventional identify-then-optimize (i.e., the thermal dynamics model is identified on historical data then used in optimization) and relaxed DFL methods in both cost savings and grid service performance, highlighting its potential for scalable, grid-aware building control.

eess.SY↗

arXiv2025

Decision-Focused Learning for Complex System Identification: HVAC Management System Application

As opposed to conventional training methods tailored to minimize a given statistical metric or task-agnostic loss (e.g., mean squared error), Decision-Focused Learning (DFL) trains machine learning models for optimal performance in downstream decision-making tools. We argue that DFL can be leveraged to learn the parameters of system dynamics, expressed as constraint of the convex optimization control policy, while the system control signal is being optimized, thus creating an end-to-end learning framework. This is particularly relevant for systems in which behavior changes once the control policy is applied, hence rendering historical data less applicable. The proposed approach can perform system identification - i.e., determine appropriate parameters for the system analytical model - and control simultaneously to ensure that the model's accuracy is focused on areas most relevant to control. Furthermore, because black-box systems are non-differentiable, we design a loss function that requires solely to measure the system response. We propose pre-training on historical data and constraint relaxation to stabilize the DFL and deal with potential infeasibilities in learning. We demonstrate the usefulness of the method on a building Heating, Ventilation, and Air Conditioning day-ahead management system for a realistic 15-zone building located in Denver, US. The results show that the conventional RC building model, with the parameters obtained from historical data using supervised learning, underestimates HVAC electrical power consumption. For our case study, the ex-post cost is on average six times higher than the expected one. Meanwhile, the same RC model with parameters obtained via DFL underestimates the ex-post cost only by 3%.

eess.SY↗