Mohamad Chehade

Mohamad Chehade, Hao Zhu

The rise of distributed energy resources (DERs) is reshaping modern distribution grids, introducing new challenges in attaining voltage stability under dynamic and decentralized operating conditions. This paper presents NEO-Grid, a unified learning-based framework for volt-var optimization (VVO) and volt-var control (VVC) that leverages neural network surrogates for power flow and deep equilibrium models (DEQs) for closed-loop control. Our method replaces traditional linear approximations with piecewise-linear ReLU networks trained to capture the nonlinear relationship between power injections and voltage magnitudes. For control, we model the recursive interaction between voltage and inverter response using DEQs, allowing direct fixed-point computation and efficient training via implicit differentiation. We evaluated NEO-Grid on the IEEE 33-bus system, demonstrating that it significantly improves voltage regulation performance compared to standard linear and heuristic baselines in both optimization and control settings. Our results establish NEO-Grid as a scalable, accurate, and interpretable solution for learning-based voltage regulation in distribution grids.

Mohamad Chehade, Sami Karaki

Sizing a residential microgrid efficiently requires solving a coupled design-and-operation problem: photovoltaic (PV) and battery capacities should be chosen in a way that reflects how the system will actually be dispatched over time. This paper proposes BOOST, or Battery-solar Ordinal Optimization Sizing Technique, which combines ordinal optimization (OO) with mixed-integer linear programming (MILP). OO is used to screen a large set of candidate battery/PV designs with a simple linear model and then re-evaluate only the most promising designs with a more accurate MILP that captures diesel commitment logic. Relative to the original short paper, this expanded manuscript retains the full methodological narrative but refreshes the quantitative section using a new synthetic benchmark dataset suite generated from the released clean reimplementation. The suite contains five yearly synthetic datasets/configurations: base, cheap battery, cheap PV, expensive diesel, and high peak tariff. On the base synthetic dataset, the best accurate design is a 500 kWh battery with 1833.3 kW of PV, achieving 13.169 c/kWh, while BOOST improves upon dynamic programming and greedy baselines. Across the full 10 x 10 design grid, the LP and MILP rankings are effectively identical (rho = 1.000), the paper-style choice of N = 90 and s = 18 recovers the global accurate optimum, and the OO-based workflow reduces runtime by 51.8% relative to exhaustive accurate evaluation on the refreshed synthetic benchmark run. Because these added datasets are synthetic, they should be read as methodological stress tests rather than as direct empirical claims about any specific real-world site.

Mohamad Chehade, Hao Zhu

Public Safety Power Shutoffs (PSPS) force rapid topology changes that can render standard operating points infeasible, requiring operators to quickly identify corrective transmission switching actions that reduce load shedding while maintaining acceptable voltage behavior. We present a verifiable, multi-stage adaptation pipeline that fine-tunes an instruction-tuned large language model (LLM) to generate \emph{open-only} corrective switching plans from compact PSPS scenario summaries under an explicit switching budget. First, supervised fine-tuning distills a DC-OPF MILP oracle into a constrained action grammar that enables reliable parsing and feasibility checks. Second, direct preference optimization refines the policy using AC-evaluated preference pairs ranked by a voltage-penalty metric, injecting voltage-awareness beyond DC imitation. Finally, best-of-$N$ selection provides an inference-time addition by choosing the best feasible candidate under the target metric. On IEEE 118-bus PSPS scenarios, fine-tuning substantially improves DC objective values versus zero-shot generation, reduces AC power-flow failure from 50\% to single digits, and improves voltage-penalty outcomes on the common-success set. Code and data-generation scripts are released to support reproducibility.

Mohamad Chehade, Soumya Suvra Ghosal, Souradip Chakraborty et al.

Aligning large language models with humans is challenging due to the inherently multifaceted nature of preference feedback. While existing approaches typically frame this as a multi-objective optimization problem, they often overlook how humans actually make decisions. Research on bounded rationality suggests that human decision making follows satisficing strategies-optimizing primary objectives while ensuring others meet acceptable thresholds. To bridge this gap and operationalize the notion of satisficing alignment, we propose SITAlign: an inference time framework that addresses the multifaceted nature of alignment by maximizing a primary objective while satisfying threshold-based constraints on secondary criteria. We provide theoretical insights by deriving sub-optimality bounds of our satisficing based inference alignment approach. We empirically validate SITAlign's performance through extensive experimentation on multiple benchmarks. For instance, on the PKU-SafeRLHF dataset with the primary objective of maximizing helpfulness while ensuring a threshold on harmlessness, SITAlign outperforms the state-of-the-art multi objective decoding strategy by a margin of 22.3% in terms of GPT-4 win-tie rate for helpfulness reward while adhering to the threshold on harmlessness.