arXiv subjects

Priya L. Donti

Publications and source records attributed to Priya L. Donti.

At least 19 recordsLinked to original sources

arXiv2026

Time to AC-t: When AC-transmission changes capacity expansion planning

Capacity expansion models often dramatically simplify the physics of how power flows through a transmission network in order to preserve computational tractability. However, with the rapid growth of electrification, large loads, renewables, and storage, increased realism in planning future grids is imperative. We extend the open-source capacity expansion model GenX with AC optimal power flow physics, and analyze generation, storage, and transmission investment decisions for 20-zone, hourly models of Germany and Texas. We compare our results against those using two common approximations, the transport model and DC optimal power flow. We find that total costs and emissions largely agree across all formulations. However, we find substantial differences in outputs related to transmission, generation siting, storage, and reactive-power support between the approximate formulations and the true AC physics formulation. We identify where AC physics changes planning decisions and discuss how to potentially capture these effects without the full computational cost.

eess.SY↗

arXiv2026

DCEmbed: Scalable Optimization over Neural Surrogates

Neural surrogates can accelerate large-scale optimization by replacing expensive or intractable model components with efficient learned approximations, but solving the resulting embedded problems can remain prohibitively costly. For instance, standard exact encodings of neural networks with ReLU activations allow the problem to be solved by mixed-integer solvers, but add large numbers of binary variables to accommodate the nonlinearity of the activations, which can render the problem computationally prohibitive. To address this, we propose DCEmbed, a heuristic for optimization problems with embedded neural surrogates that leverages the difference-of-convex (DC) representation of the network and avoids adding activation binaries. Exploiting shared structure within the DC representation of a ReLU network, we derive a reduced-size, exact formulation for its convex components that can be embedded in optimization problems using just two linear inequalities and one continuous auxiliary variable per hidden neuron. Using this, the problem is solved via an iterative penalty convex-concave procedure, where only the concave portions of the neural terms are approximated at each stage. The original objective, constraints, and any discrete decisions are retained, allowing standard convex or mixed-integer optimization solvers to optimize the host and surrogate jointly at each iteration. In experiments on quadratic programs, mixed-integer resource allocation, and neural two-stage stochastic programming, our method demonstrates much faster progress toward high-quality feasible solutions than approaches using exact mixed-integer embeddings. In particular, DCEmbed achieves $4\times$ lower normalized primal integral than the best exact baseline on the resource allocation problem, while in two-stage stochastic programming it reaches the global surrogate optimum $\sim 5\times$ faster than Gurobi ML.

math.OC↗

arXiv2026

Scalable Self-Supervised Learning for Multiphase AC-OPF in Distribution Systems with Topology Reconfiguration

The proliferation of distributed energy resources (DERs) in distribution grids enables the active coordination of these assets to reduce costs and enable cleaner operations. Realizing this potential requires solving multiphase AC optimal power flow (AC-OPF) quickly across varying loads, DER availabilities, and topology reconfigurations, at much greater speed and scale than conventional nonlinear solvers. Learning-based surrogates can offer millisecond inference, yet existing methods target largely balanced transmission systems and do not scale to the multiphase, unbalanced, and reconfigurable nature of distribution feeders at utility scale. We present the Penalty + Sequential Linearized Feasibility Seeking (SLFS) algorithm, a self-supervised learning framework for multiphase distribution AC-OPF under switch-induced topology changes. Penalty+SLFS requires no labeled optimal solutions and trains directly from the AC-OPF objective and constraints through a differentiable fixed-point power flow solver, avoiding expensive label generation and admitting robust training procedures. Topology changes are handled efficiently using Sherman-Morrison-Woodbury updates of the admittance-matrix inverse, while an M-step Jacobian approximation accelerates differentiation through the power flow solver. At inference, SLFS repairs any infeasible predictions, providing feasibility guarantees with low computational overhead. On IEEE feeders ranging from 13 to 8,500 nodes, Penalty+SLFS achieves negligible optimality gaps and near-zero constraint violations, delivers up to three orders of magnitude speedups over IPOPT, and remains robust under large distributional shifts, demonstrating a viable path toward real-time, topology-aware AC-OPF for large-scale distribution grids.

eess.SY↗

arXiv2026

Modeling of Network Constraints in Large-scale Capacity Expansion Optimization of Power Grids

Capacity expansion modeling plays a critical role in optimizing the deployment of new generation, storage, and transmission, typically at national and regional levels. To support long-term planning, these models consider a large set of energy technologies and policies, along with decades of weather and demand data. Realistic capacity expansion models thus become high-dimensional optimization problems, with hundreds of millions of variables and constraints, which are challenging to solve. A common strategy to address this complexity is to omit non-linear, non-convex AC optimal power flow (ACOPF) constraints and instead use linearized power balance equations or transport formulations. While these simplifications improve tractability, they limit our understanding of how power flow and the physical properties of power networks impact investment decisions across generation, storage, and transmission infrastructure. This paper addresses this gap by extending the GenX capacity expansion model to incorporate fixed point theorem-based network constraints. These embed ACOPF-based considerations while maintaining the tractability of the planning model, nearly preserving the dimensionality of the transport formulation and incurring only modest runtime increases. This approach is much cheaper than embedding ACOPF directly, making it appropriate for large-scale capacity planning problems. We compare our approach to the original transport-based GenX model as well as a non-linear, non-convex version that incorporates the full ACOPF constraints, for a case study of the ISO New England grid.

math.OC↗

arXiv2026

Physics-Informed Graph Neural Networks for Robust AC-Optimal Power Flow

We present PINCO, an unsupervised learning framework that integrates Graph Neural Networks with physics-informed neural networks for AC optimal power flow (AC-OPF) solutions. Unlike state-of-the-art unsupervised methods that require prescreened datasets containing only feasible instances, our approach operates on unfiltered data, including ill-conditioned cases. The framework addresses two critical gaps in the literature: (1) robustness to topology changes up to N-2 contingencies, and (2) detecting optimal power flow instances that are infeasible without relying on traditional solvers for data filtering. PINCO embeds physical laws into the learning process via augmented Lagrangian multipliers. In addition, it introduces a clustering branch with learnable centroids that automatically separate feasible from infeasible solutions based on constraint-violation patterns. We evaluate the framework across systems of varying complexity, including the IEEE 30-bus, IEEE 57-bus, and Swiss transmission networks, demonstrating scalability and robustness under diverse loading conditions and topology variations. Benchmarking against DeepOPF-FT and the IPOPT solver shows that PINCO achieves comparable constraint satisfaction while delivering two to three orders of magnitude computational speedup compared to IPOPT and lower operational costs across all configurations.

eess.SY↗

arXiv2026

Cheap Thrills: Effective Amortized Optimization Using Inexpensive Labels

To scale optimization and simulation, prior work has explored training machine-learning surrogates that map problem parameters to solutions inexpensively at inference time. Unfortunately, commonly used approaches, including supervised and self-supervised learning with either soft or hard feasibility enforcement, face inherent challenges such as reliance on expensive high-quality labels or difficult optimization landscapes. To address their trade-offs, we propose a novel framework that collects "cheap" imperfect labels, performs supervised model pretraining with a merit loss-based termination scheme, and finally refines the model through self-supervised learning to improve final performance. Empirical validation across challenging domains -- including nonconvex constrained optimization, power-grid operation, and stiff dynamical systems -- shows that this three-stage strategy yields faster convergence; improved accuracy, feasibility, and optimality; and up to 59x reductions in total offline computational cost. We further analyze why and when our framework improves surrogate model training, finding that (i) merit loss is an informative signal and (ii) only small numbers of cheap, inexact labels are needed to place the model in a favorable regime for self-supervised learning.

cs.LG↗

arXiv2026

Improving Feasibility via Fast Autoencoder-Based Projections

Enforcing complex (e.g., nonconvex) operational constraints is a critical challenge in real-world learning and control systems. However, existing methods struggle to efficiently enforce general classes of constraints. To address this, we propose a novel data-driven amortized approach that uses a trained autoencoder as an approximate projector to provide fast corrections to infeasible predictions. Specifically, we train an autoencoder using an adversarial objective to learn a structured, convex latent representation of the feasible set. This enables rapid correction of neural network outputs by projecting their associated latent representations onto a simple convex shape before decoding into the original feasible set. We test our approach on a diverse suite of constrained optimization and reinforcement learning problems with challenging nonconvex constraints. Results show that our method effectively enforces constraints at a low computational cost, offering a practical alternative to expensive feasibility correction techniques based on traditional solvers.

cs.LG↗

arXiv2025

Application-Driven Innovation in Machine Learning

In this position paper, we argue that application-driven research has been systemically under-valued in the machine learning community. As applications of machine learning proliferate, innovative algorithms inspired by specific real-world challenges have become increasingly important. Such work offers the potential for significant impact not merely in domains of application but also in machine learning itself. In this paper, we describe the paradigm of application-driven research in machine learning, contrasting it with the more standard paradigm of methods-driven research. We illustrate the benefits of application-driven machine learning and how this approach can productively synergize with methods-driven work. Despite these benefits, we find that reviewing, hiring, and teaching practices in machine learning often hold back application-driven innovation. We outline how these processes may be improved.

cs.LG↗

arXiv2025

Strategic bid response under automated market power mitigation in electricity markets

In auction markets that are prone to market power abuse, preventive mitigation of bid prices can be applied through automated mitigation procedures (AMP). Despite the widespread application of AMP in US electricity markets, there exists scarce evidence on how firms strategically react to such price-cap-and-penalty regulation: when the price cap rarely leads to penalty mitigation, it is difficult to distinguish whether AMP are an effective deterrent or simply too lax. We investigate their impact on the bids of generation firms, using 2019 data from the New York and New England electricity markets (NYISO, ISO-NE). We employ a regression discontinuity design, which exploits the fact that the price cap with penalty is only activated when a structural index (e.g., congestion, pivotality) exceeds a certain cutoff. By estimating the Local Average Treatment Effect (LATE) of screening activation, we can causally identify successful deterrence of anti-competitive behavior. Around 30-40% of the analyzed bidders per market exhibit a significant strategic response - corresponding to a decrease in maximum bid prices of 4-10 $/MWh to avoid the penalty. However, there is significant heterogeneity between firms, and the regulatory impact on the overall market is not statistically detectable, suggesting lax mitigation thresholds. Using a merit-order simulation, we estimate the welfare impact of more stringent thresholds to lie between 350 and 980 thousand dollars of increased buyer surplus per mitigated hour, with the associated number of mitigated hours being below 33 hours/year. Our results motivate the empirical calibration of mitigation thresholds to improve the efficiency of AMP regulation.

econ.GN↗

arXiv2025

FSNet: Feasibility-Seeking Neural Network for Constrained Optimization with Guarantees

Efficiently solving constrained optimization problems is crucial for numerous real-world applications, yet traditional solvers are often computationally prohibitive for real-time use. Machine learning-based approaches have emerged as a promising alternative to provide approximate solutions at faster speeds, but they struggle to strictly enforce constraints, leading to infeasible solutions in practice. To address this, we propose the Feasibility-Seeking Neural Network (FSNet), which integrates a feasibility-seeking step directly into its solution procedure to ensure constraint satisfaction. This feasibility-seeking step solves an unconstrained optimization problem that minimizes constraint violations in a differentiable manner, enabling end-to-end training and providing guarantees on feasibility and convergence. Our experiments across a range of different optimization problems, including both smooth/nonsmooth and convex/nonconvex problems, demonstrate that FSNet can provide feasible solutions with solution quality comparable to (or in some cases better than) traditional solvers, at significantly faster speeds.

cs.LG↗

arXiv2025

RL2Grid: Benchmarking Reinforcement Learning in Power Grid Operations

Reinforcement learning (RL) can provide adaptive and scalable controllers essential for power grid decarbonization. However, RL methods struggle with power grids' complex dynamics, long-horizon goals, and hard physical constraints. For these reasons, we present RL2Grid, a benchmark designed in collaboration with power system operators to accelerate progress in grid control and foster RL maturity. Built on RTE France's power simulation framework, RL2Grid standardizes tasks, state and action spaces, and reward structures for a systematic evaluation and comparison of RL algorithms. Moreover, we integrate operational heuristics and design safety constraints based on human expertise to ensure alignment with physical requirements. By establishing reference performance metrics for classic RL baselines on RL2Grid's tasks, we highlight the need for novel methods capable of handling real systems and discuss future directions for RL-based grid control.

cs.LG↗

arXiv2025

Improving Policy Optimization via $\varepsilon$-Retrain

We present $\varepsilon$-retrain, an exploration strategy encouraging a behavioral preference while optimizing policies with monotonic improvement guarantees. To this end, we introduce an iterative procedure for collecting retrain areas -- parts of the state space where an agent did not satisfy the behavioral preference. Our method switches between the typical uniform restart state distribution and the retrain areas using a decaying factor $\varepsilon$, allowing agents to retrain on situations where they violated the preference. We also employ formal verification of neural networks to provably quantify the degree to which agents adhere to these behavioral preferences. Experiments over hundreds of seeds across locomotion, power network, and navigation tasks show that our method yields agents that exhibit significant performance and sample efficiency improvements.

cs.AI↗

arXiv2024

Defining 'Good': Evaluation Framework for Synthetic Smart Meter Data

Access to granular demand data is essential for the net zero transition; it allows for accurate profiling and active demand management as our reliance on variable renewable generation increases. However, public release of this data is often impossible due to privacy concerns. Good quality synthetic data can circumnavigate this issue. Despite significant research on generating synthetic smart meter data, there is still insufficient work on creating a consistent evaluation framework. In this paper, we investigate how common frameworks used by other industries leveraging synthetic data, can be applied to synthetic smart meter data, such as fidelity, utility and privacy. We also recommend specific metrics to ensure that defining aspects of smart meter data are preserved and test the extent to which privacy can be protected using differential privacy. We show that standard privacy attack methods like reconstruction or membership inference attacks are inadequate for assessing privacy risks of smart meter datasets. We propose an improved method by injecting training data with implausible outliers, then launching privacy attacks directly on these outliers. The choice of $ε$ (a metric of privacy loss) significantly impacts privacy risk, highlighting the necessity of performing these explicit privacy tests when making trade-offs between fidelity and privacy.

cs.LG↗

arXiv2023

Proceedings of AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges

Climate change is one of the most pressing challenges of our time, requiring rapid action across society. As artificial intelligence tools (AI) are rapidly deployed, it is therefore crucial to understand how they will impact climate action. On the one hand, AI can support applications in climate change mitigation (reducing or preventing greenhouse gas emissions), adaptation (preparing for the effects of a changing climate), and climate science. These applications have implications in areas ranging as widely as energy, agriculture, and finance. At the same time, AI is used in many ways that hinder climate action (e.g., by accelerating the use of greenhouse gas-emitting fossil fuels). In addition, AI technologies have a carbon and energy footprint themselves. This symposium brought together participants from across academia, industry, government, and civil society to explore these intersections of AI with climate change, as well as how each of these sectors can contribute to solutions.

cs.AI↗

arXiv2021

Adversarially Robust Learning for Security-Constrained Optimal Power Flow

In recent years, the ML community has seen surges of interest in both adversarially robust learning and implicit layers, but connections between these two areas have seldom been explored. In this work, we combine innovations from these areas to tackle the problem of N-k security-constrained optimal power flow (SCOPF). N-k SCOPF is a core problem for the operation of electrical grids, and aims to schedule power generation in a manner that is robust to potentially k simultaneous equipment outages. Inspired by methods in adversarially robust training, we frame N-k SCOPF as a minimax optimization problem - viewing power generation settings as adjustable parameters and equipment outages as (adversarial) attacks - and solve this problem via gradient-based techniques. The loss function of this minimax problem involves resolving implicit equations representing grid physics and operational decisions, which we differentiate through via the implicit function theorem. We demonstrate the efficacy of our framework in solving N-3 SCOPF, which has traditionally been considered as prohibitively expensive to solve given that the problem size depends combinatorially on the number of potential outages.

math.OC↗

arXiv2021

DC3: A learning method for optimization with hard constraints

Large optimization problems with hard constraints arise in many settings, yet classical solvers are often prohibitively slow, motivating the use of deep networks as cheap "approximate solvers." Unfortunately, naive deep learning approaches typically cannot enforce the hard constraints of such problems, leading to infeasible solutions. In this work, we present Deep Constraint Completion and Correction (DC3), an algorithm to address this challenge. Specifically, this method enforces feasibility via a differentiable procedure, which implicitly completes partial solutions to satisfy equality constraints and unrolls gradient-based corrections to satisfy inequality constraints. We demonstrate the effectiveness of DC3 in both synthetic optimization tasks and the real-world setting of AC optimal power flow, where hard constraints encode the physics of the electrical grid. In both cases, DC3 achieves near-optimal objective values while preserving feasibility.

cs.LG↗

arXiv2021

Enforcing robust control guarantees within neural network policies

When designing controllers for safety-critical systems, practitioners often face a challenging tradeoff between robustness and performance. While robust control methods provide rigorous guarantees on system stability under certain worst-case disturbances, they often yield simple controllers that perform poorly in the average (non-worst) case. In contrast, nonlinear control methods trained using deep learning have achieved state-of-the-art performance on many control tasks, but often lack robustness guarantees. In this paper, we propose a technique that combines the strengths of these two approaches: constructing a generic nonlinear control policy class, parameterized by neural networks, that nonetheless enforces the same provable robustness criteria as robust control. Specifically, our approach entails integrating custom convex-optimization-based projection layers into a neural network-based policy. We demonstrate the power of this approach on several domains, improving in average-case performance over existing robust control methods and in worst-case stability over (non-robust) deep RL methods.

cs.LG↗

arXiv2019

Tackling Climate Change with Machine Learning

Climate change is one of the greatest challenges facing humanity, and we, as machine learning experts, may wonder how we can help. Here we describe how machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. From smart grids to disaster management, we identify high impact problems where existing gaps can be filled by machine learning, in collaboration with other fields. Our recommendations encompass exciting research questions as well as promising business opportunities. We call on the machine learning community to join the global effort against climate change.

cs.CY↗