Cosmin G. Petra
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Cosmin G. Petra has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
- Jingyi Wang
Center for Applied Scientific Computing Lawrence Livermore National Laboratory Livermore CA
5 shared publications - Emil Mihai Constantinescu
Mathematics and Computer Science Division Argonne National Laboratory 60439 Lemont IL USA
3 shared publications - Ignacio Aravena
Lawrence Livermore National Laboratory 94550 Livermore CA
3 shared publications - Johannes J. Brust
Argonne National Laboratory
3 shared publications - Noemi Petra
Applied Mathematics, School of Natural Sciences, University of California, Merced, CA 95343, United States of America
3 shared publications - Roummel F. Marcia
Department of Applied Mathematics, University of California Merced, Merced, CA, USA
3 shared publications
Latest Publications
- 17 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Systems
- Vol. PP (08858950) , 1-14
Multi-Fidelity Dynamic Line Rating Fusion for System Load Margin Enhancement with Large-Scale Offshore Wind Generations
This paper proposes a multi-fidelity dynamic line rating (DLR) fusion method for power system load margin enhancement with large-scale offshore wind integration. First, multi-fidelity DLR data from diverse sensors are fused using multi-fidelity deep sigma point processes (MFDSPP), a probabilistic nonlinear learning model that integrates heterogeneous data sources to estimate the minimum line rating and quantify its uncertainty. Conformal prediction (CP), a distribution free uncertainty calibration technique that refines prediction intervals using empirical residuals, is developed to refine the uncertainty intervals of DLRs. Then, continuation power flow (CPF) analysis is used to calculate the system's maximum load margin for numerous N − k contingencies. Sampling these scenarios generates a large dataset to train the deep sigma point processes (DSPP) model, establishing a direct nonlinear relationship between operational data and load margin. Finally, DLR based line capacity constraints and DSPP-derived load-margin constraints are integrated into the real-time control framework to improve system stability. Case studies on the Northeast Power Coordinating Council (NPCC) system, incorporating real DLR sensor measurements, authentic N −k contingency datasets, and four real large-scale offshore wind farms, demonstrate that the proposed method effectively maintains a load margin greater than 10% of the base load, unlocks an additional 2,189.7 MWh of clean energy, reduces operating costs by approximately 2.34%, and achieves more than a 2,000× speedup compared with classical CPF-based sensitivity methods.
- 17 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Systems
- Vol. PP (08858950) , 1-14
Multi-Fidelity Dynamic Line Rating Fusion for System Load Margin Enhancement with Large-Scale Offshore Wind Generations
This paper proposes a multi-fidelity dynamic line rating (DLR) fusion method for power system load margin enhancement with large-scale offshore wind integration. First, multi-fidelity DLR data from diverse sensors are fused using multi-fidelity deep sigma point processes (MFDSPP), a probabilistic nonlinear learning model that integrates heterogeneous data sources to estimate the minimum line rating and quantify its uncertainty. Conformal prediction (CP), a distribution free uncertainty calibration technique that refines prediction intervals using empirical residuals, is developed to refine the uncertainty intervals of DLRs. Then, continuation power flow (CPF) analysis is used to calculate the system's maximum load margin for numerous N − k contingencies. Sampling these scenarios generates a large dataset to train the deep sigma point processes (DSPP) model, establishing a direct nonlinear relationship between operational data and load margin. Finally, DLR based line capacity constraints and DSPP-derived load-margin constraints are integrated into the real-time control framework to improve system stability. Case studies on the Northeast Power Coordinating Council (NPCC) system, incorporating real DLR sensor measurements, authentic N −k contingency datasets, and four real large-scale offshore wind farms, demonstrate that the proposed method effectively maintains a load margin greater than 10% of the base load, unlocks an additional 2,189.7 MWh of clean energy, reduces operating costs by approximately 2.34%, and achieves more than a 2,000× speedup compared with classical CPF-based sensitivity methods.
- 8 April 2026
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Scientific Computing
- Vol. 48 (2) , A828-A852
Algebraic Multigrid with Filtering: An Efficient Preconditioner for Interior Point Methods in Large-Scale Contact Mechanics Optimization
Large-scale contact mechanics simulations are crucial in many engineering fields such as structural design and manufacturing. In the frictionless case, contact can be modeled by minimizin...
- 8 April 2026
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Scientific Computing
- Vol. 48 (2) , A828-A852
Algebraic Multigrid with Filtering: An Efficient Preconditioner for Interior Point Methods in Large-Scale Contact Mechanics Optimization
Large-scale contact mechanics simulations are crucial in many engineering fields such as structural design and manufacturing. In the frictionless case, contact can be modeled by minimizin...
- 15 October 2025
- Published by Elsevier in Computers & Chemical Engineering
- Vol. 205
A grid-scale study of demand bidding by large industrial users
- 15 October 2025
- Published by Elsevier in Computers & Chemical Engineering
- Vol. 205
A grid-scale study of demand bidding by large industrial users
- Open Access
A Constrained Multi-Fidelity Bayesian Optimization Method
Recently, multi-fidelity Bayesian optimization (MFBO) has been successfully applied to many engineering design optimization problems, where the cost of high-fidelity simulations and experiments can be prohibitive. However, challenges remain for constrained optimization problems using the MFBO framework, particularly in efficiently identifying the feasible region defined by the constraints. In this paper, we propose a constrained multi-fidelity Bayesian optimization (CMFBO) method with novel acquisition functions. Specifically, we design efficient acquisition functions that 1) have analytically closed-form expressions; 2) are straightforward to implement; and 3) do not require feasible initial samples, an important feature often missing in commonly used acquisition functions such as expected constrained improvement (ECI). We demonstrate the effectiveness of our algorithms on synthetic test problems using different combinations of acquisition functions. Then, we apply the proposed method to a data-driven inertial confinement fusion (ICF) design problem, and a high-current joint design problem using finite element simulations with computational contact mechanics.
Open AccessA Constrained Multi-Fidelity Bayesian Optimization Method
Recently, multi-fidelity Bayesian optimization (MFBO) has been successfully applied to many engineering design optimization problems, where the cost of high-fidelity simulations and experiments can be prohibitive. However, challenges remain for constrained optimization problems using the MFBO framework, particularly in efficiently identifying the feasible region defined by the constraints. In this paper, we propose a constrained multi-fidelity Bayesian optimization (CMFBO) method with novel acquisition functions. Specifically, we design efficient acquisition functions that 1) have analytically closed-form expressions; 2) are straightforward to implement; and 3) do not require feasible initial samples, an important feature often missing in commonly used acquisition functions such as expected constrained improvement (ECI). We demonstrate the effectiveness of our algorithms on synthetic test problems using different combinations of acquisition functions. Then, we apply the proposed method to a data-driven inertial confinement fusion (ICF) design problem, and a high-current joint design problem using finite element simulations with computational contact mechanics.
- Open Access
Bayesian Optimization with Expected Improvement: No Regret and the Choice of Incumbent
Expected improvement (EI) is one of the most widely used acquisition functions in Bayesian optimization (BO). Despite its proven empirical success in applications, the cumulative regret upper bound of EI remains an open question. In this paper, we analyze the classic noisy Gaussian process expected improvement (GP-EI) algorithm. We consider the Bayesian setting, where the objective is a sample from a GP. Three commonly used incumbents, namely the best posterior mean incumbent (BPMI), the best sampled posterior mean incumbent (BSPMI), and the best observation incumbent (BOI) are considered as the choices of the current best value in GP-EI. We present for the first time the cumulative regret upper bounds of GP-EI with BPMI and BSPMI. Importantly, we show that in both cases, GP-EI is a no-regret algorithm for both squared exponential (SE) and Matérn kernels. Further, we present for the first time that GP-EI with BOI either achieves a sublinear cumulative regret upper bound or has a fast converging noisy simple regret bound for SE and Matérn kernels. Our results provide theoretical guidance to the choice of incumbent when practitioners apply GP-EI in the noisy setting. Numerical experiments are conducted to validate our findings.
Open AccessBayesian Optimization with Expected Improvement: No Regret and the Choice of Incumbent
Expected improvement (EI) is one of the most widely used acquisition functions in Bayesian optimization (BO). Despite its proven empirical success in applications, the cumulative regret upper bound of EI remains an open question. In this paper, we analyze the classic noisy Gaussian process expected improvement (GP-EI) algorithm. We consider the Bayesian setting, where the objective is a sample from a GP. Three commonly used incumbents, namely the best posterior mean incumbent (BPMI), the best sampled posterior mean incumbent (BSPMI), and the best observation incumbent (BOI) are considered as the choices of the current best value in GP-EI. We present for the first time the cumulative regret upper bounds of GP-EI with BPMI and BSPMI. Importantly, we show that in both cases, GP-EI is a no-regret algorithm for both squared exponential (SE) and Matérn kernels. Further, we present for the first time that GP-EI with BOI either achieves a sublinear cumulative regret upper bound or has a fast converging noisy simple regret bound for SE and Matérn kernels. Our results provide theoretical guidance to the choice of incumbent when practitioners apply GP-EI in the noisy setting. Numerical experiments are conducted to validate our findings.
- Open Access
- 29 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computing in Science & Engineering
- Vol. 26 (2) , 25-32
Advanced Computing is at the Forefront of a New “Moonshot” Revolutionizing the North American Power Grid
In the 50+ years since the first humans landed on the moon, computing has grown at breakneck speed. We are faced with another challenge that is just as daunting, and just as important to overcome—modernizing the North American electric power grid—and high-performance computing (HPC) systems with specialized software will be an important element in rising to this challenge. We describe at a high level how software developed in the ExaSGD project addresses this “moonshot” goal by utilizing exascale computing and a novel high performance solver software stack to support the mission of decarbonizing power grid operations in an environment of uncertain weather and climate. To reach the exascale benchmark the team has made a number of first-of-their-kind innovations, including novel method for stochastic optimization, fine grained parallel methods for modeling power systems, and GPU resident sparse numerical linear solvers.
- 29 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computing in Science & Engineering
- Vol. 26 (2) , 25-32
Open AccessAdvanced Computing is at the Forefront of a New “Moonshot” Revolutionizing the North American Power Grid
In the 50+ years since the first humans landed on the moon, computing has grown at breakneck speed. We are faced with another challenge that is just as daunting, and just as important to overcome—modernizing the North American electric power grid—and high-performance computing (HPC) systems with specialized software will be an important element in rising to this challenge. We describe at a high level how software developed in the ExaSGD project addresses this “moonshot” goal by utilizing exascale computing and a novel high performance solver software stack to support the mission of decarbonizing power grid operations in an environment of uncertain weather and climate. To reach the exascale benchmark the team has made a number of first-of-their-kind innovations, including novel method for stochastic optimization, fine grained parallel methods for modeling power systems, and GPU resident sparse numerical linear solvers.
- Open Access
Constrained Bayesian optimization with merit functions
Bayesian optimization is a powerful optimization tool for problems where native first-order derivatives are unavailable. Recently, constrained Bayesian optimization (CBO) has been applied to many engineering applications where constraints are essential. However, several obstacles remain with current CBO algorithms that could prevent a wider adoption. We propose CBO algorithms using merit functions, such as the penalty merit function, in acquisition functions, inspired by nonlinear optimization methods, e.g., sequential quadratic programming. Merit functions measure the potential progress of both the objective and constraint functions, thus increasing algorithmic efficiency and allowing infeasible initial samples. The acquisition functions with merit functions are relaxed to have closed forms, making its implementation readily available wherever Bayesian optimization is. We further propose a unified CBO algorithm that can be seen as extension to the popular expected constrained improvement (ECI) approach. We demonstrate the effectiveness and efficiency of the proposed algorithms through numerical experiments on synthetic problems and a practical data-driven engineering design problem in the field of plasma physics.
Open AccessConstrained Bayesian optimization with merit functions
Bayesian optimization is a powerful optimization tool for problems where native first-order derivatives are unavailable. Recently, constrained Bayesian optimization (CBO) has been applied to many engineering applications where constraints are essential. However, several obstacles remain with current CBO algorithms that could prevent a wider adoption. We propose CBO algorithms using merit functions, such as the penalty merit function, in acquisition functions, inspired by nonlinear optimization methods, e.g., sequential quadratic programming. Merit functions measure the potential progress of both the objective and constraint functions, thus increasing algorithmic efficiency and allowing infeasible initial samples. The acquisition functions with merit functions are relaxed to have closed forms, making its implementation readily available wherever Bayesian optimization is. We further propose a unified CBO algorithm that can be seen as extension to the popular expected constrained improvement (ECI) approach. We demonstrate the effectiveness and efficiency of the proposed algorithms through numerical experiments on synthetic problems and a practical data-driven engineering design problem in the field of plasma physics.
- 1 November 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (6) , 1997-2014
Recent Developments in Security-Constrained AC Optimal Power Flow: Overview of Challenge 1 in the ARPA-E Grid Optimization Competition
Intro to the ARPA-E Grid Optimization Competition In “Recent Developments in Security-Constrained AC Optimal Power Flow: Overview of Challenge 1 in the ARPA-E Grid Optimization Competition,” we review the state of the art in practical algorithms for scheduling power-systems operations in the short term and the results of the recent competition organized by the U.S. Advanced Research Projects Agency–Energy. We explain the mixed-integer nonlinear formulation used in the competition for nonspecialists in electrical engineering, the context and organization of the competition, and the performance of competitors. We find that the collective approaches and results of competitors provide support for efforts to move nonlinear optimization techniques into industrial applications, as they have proven to be a robust and efficient alternative to current linear approximation techniques.
- 1 November 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (6) , 1997-2014
Recent Developments in Security-Constrained AC Optimal Power Flow: Overview of Challenge 1 in the ARPA-E Grid Optimization Competition
Intro to the ARPA-E Grid Optimization Competition In “Recent Developments in Security-Constrained AC Optimal Power Flow: Overview of Challenge 1 in the ARPA-E Grid Optimization Competition,” we review the state of the art in practical algorithms for scheduling power-systems operations in the short term and the results of the recent competition organized by the U.S. Advanced Research Projects Agency–Energy. We explain the mixed-integer nonlinear formulation used in the competition for nonspecialists in electrical engineering, the context and organization of the competition, and the performance of competitors. We find that the collective approaches and results of competitors provide support for efforts to move nonlinear optimization techniques into industrial applications, as they have proven to be a robust and efficient alternative to current linear approximation techniques.
- 1 November 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (6) , 2015-2030
A Surrogate-Based Asynchronous Decomposition Technique for Realistic Security-Constrained Optimal Power Flow Problems
Solving realistic security-constrained optimal power flow problems In “A surrogate-based asynchronous decomposition technique for realistic security-constrained optimal power flow problems,” we propose a new algorithm for solving a classical problem in power grid operations: the security-constrained optimal power flow, considering its nonlinearities and realistic transitions between nominal and emergency post-contingency operations. Solving security-constrained optimal power flow problems accurately is a critical function, upon which depends the reliability, security, and efficiency of power systems as well as the correct functioning of other critical infrastructure dependent on electricity. The proposed algorithm was extensively tested against many state-of-the-art approaches using realistic and real instances in the ARPA-E Grid Optimization Competition Challenge 1, where it found the best-known solution for 58% of the instances, attained an average gap of less than 0.2%, and obtained the best overall scores, thereby winning all divisions of Challenge 1 with a very strong first place.
- 1 November 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (6) , 2015-2030
A Surrogate-Based Asynchronous Decomposition Technique for Realistic Security-Constrained Optimal Power Flow Problems
Solving realistic security-constrained optimal power flow problems In “A surrogate-based asynchronous decomposition technique for realistic security-constrained optimal power flow problems,” we propose a new algorithm for solving a classical problem in power grid operations: the security-constrained optimal power flow, considering its nonlinearities and realistic transitions between nominal and emergency post-contingency operations. Solving security-constrained optimal power flow problems accurately is a critical function, upon which depends the reliability, security, and efficiency of power systems as well as the correct functioning of other critical infrastructure dependent on electricity. The proposed algorithm was extensively tested against many state-of-the-art approaches using realistic and real instances in the ARPA-E Grid Optimization Competition Challenge 1, where it found the best-known solution for 58% of the instances, attained an average gap of less than 0.2%, and obtained the best overall scores, thereby winning all divisions of Challenge 1 with a very strong first place.
- Open Access
Local convergence of a sequential quadratic programming method for a class of nonsmooth nonconvex objectives
A sequential quadratic programming (SQP) algorithm is designed for nonsmooth optimization problems with upper-C^2 objective functions. Upper-C^2 functions are locally equivalent to difference-of-convex (DC) functions with smooth convex parts. They arise naturally in many applications such as certain classes of solutions to parametric optimization problems, e.g., recourse of stochastic programming, and projection onto closed sets. The proposed algorithm conducts line search and adopts an exact penalty merit function. The potential inconsistency due to the linearization of constraints are addressed through relaxation, similar to that of Sl_1QP. We show that the algorithm is globally convergent under reasonable assumptions. Moreover, we study the local convergence behavior of the algorithm under additional assumptions of Kurdyka-Łojasiewicz (KL) properties, which have been applied to many nonsmooth optimization problems. Due to the nonconvex nature of the problems, a special potential function is used to analyze local convergence. We show that under acceptable assumptions, upper bounds on local convergence can be proven. Additionally, we show that for a large number of optimization problems with upper-C^2 objectives, their corresponding potential functions are indeed KL functions. Numerical experiment is performed with a power grid optimization problem that is consistent with the assumptions and analysis in this paper.
Open AccessLocal convergence of a sequential quadratic programming method for a class of nonsmooth nonconvex objectives
A sequential quadratic programming (SQP) algorithm is designed for nonsmooth optimization problems with upper-C^2 objective functions. Upper-C^2 functions are locally equivalent to difference-of-convex (DC) functions with smooth convex parts. They arise naturally in many applications such as certain classes of solutions to parametric optimization problems, e.g., recourse of stochastic programming, and projection onto closed sets. The proposed algorithm conducts line search and adopts an exact penalty merit function. The potential inconsistency due to the linearization of constraints are addressed through relaxation, similar to that of Sl_1QP. We show that the algorithm is globally convergent under reasonable assumptions. Moreover, we study the local convergence behavior of the algorithm under additional assumptions of Kurdyka-Łojasiewicz (KL) properties, which have been applied to many nonsmooth optimization problems. Due to the nonconvex nature of the problems, a special potential function is used to analyze local convergence. We show that under acceptable assumptions, upper bounds on local convergence can be proven. Additionally, we show that for a large number of optimization problems with upper-C^2 objectives, their corresponding potential functions are indeed KL functions. Numerical experiment is performed with a power grid optimization problem that is consistent with the assumptions and analysis in this paper.