My second research stay at NAG (Oxford-UK)

  My secondment in NAG took place online in November-December 2021 in collaboration with Jan Fiala and Shuanghua Bai. As an introduction, NAG provides industry-leading numerical software and technical services to banking and finance, energy, engineering, and market research, as well as academic and government institutions. NAG also offers Automatic Differentiation, Machine Learning, and Mathematical Optimization products, as well as world-class technical consultancy across HPC and Cloud HPC, code porting and optimization, and other areas of numerical computing. Founded over 50 years ago from a multi-university venture, NAG is headquartered in Oxford, UK, with offices in the UK, US, EU, and Asia. At FAU (my host university), my supervisor (Prof. Michael Stingl) and I could design a Primal-Dual Penalty/Barrier Multiplier Augmented Lagrangian (PBM-AL) algorithm for the solution of Semidefinite Programs (SDP), the efficiency of the overall algorithm is demonstrated by numerical experiments for different class of SDPs, including examples from SDPLIB, medium to significant graph problems as well as truss topology optimization problems. The purpose of this secondment was to computationally study and explore how to effectively solve a particular class of SDP problems, including SDP relaxations coming from optimal power flow (OPF) problems, by using our developed algorithm at FAU. The classical OPF problem is a nonconvex NLP.

On the other hand, the SDP belongs to convex optimization and can guarantee optimal global solutions using a desired algorithm. Hence, it is worth studying how to correctly reformulate the classical OPF to an SDP model and benefits from the SDP techniques. Related literature, as well as a package called "MATPOWER"-an open source Matlab-language M-files for solving steady-state power system simulation and optimization problems such as power flow 3 - has been introduced by NAG, as it was a new subject for me in our weekly online meetings with Shuanghau Bai, we mainly discussed on the structure and modeling of the SDP relaxation for these specific problems. In this open-source MATPOWER package, another software package for optimization modeling tool YALMIP and a semidefinite programming solver compatible with YALMIP, such as SEDUMI or SDPT3, are required. The plan was to make the required changes in modeling OPF problems in order to be able to plug in the Primal-Dual PBM-AL solver instead of so-called solvers, so I was mainly focusing on both solvers to see how our algorithm could work efficiently or if it could be improved for these specific problems. As the duration of this secondment needed to be increased for this goal, it is an ongoing collaboration and still needs more time for the numerical results to be finalized.

          Newsletter April 2022.pdf - Google Drive 





 

Comments

Post a Comment

Popular posts from this blog

Summary of my research's theoretical part

Image
During the last year of my study, we have been exploring the convergence analysis of the developed algorithm (PBM-AL) . We developed a dual convergence analysis for the penalty barrier multiplier augmented Lagrangian (PBM-AL) method using a hyperbolic penalty function. Our analysis showed that PBM-AL can be characterized as a proximal point algorithm based on Fenchel conjugates. This allows for a more efficient scheme without sacrificing convergence results. Notably, our analysis relaxes the assumption of exact subproblem solutions, introducing an early stopping condition based on Fenchel duality. We also proposed a primal-dual update scheme ( PDAL ) to ensure dually feasible iterates and allowed controlled violation of equality constraints for the use of iterative solvers. A relaxation of Fenchel duality using an exact penalty function was utilized for the early stopping criterion. Below are the key literature sources that were utilized for the analysis and derivation of the convergen
Read more

My third (final) research stay at Tilburg University

Image
Fortunately, as COVID-19  😫  restrictions gradually lifted, I received confirmation that I could undertake my research stay on-site at Tilburg University in the Netherlands, commencing on April 1, 2022, and continuing until the end of May 2022 😇 (a duration of 2 months). My initial impression of the Netherlands was overwhelmingly positive 😍. The entire country was teeming with bicycles, which was entirely logical given its remarkably flat terrain. Moreover, I was fortunate to be there during the most enchanting time of the year—springtime  ⚘😍 !  Undoubtedly, everyone is familiar with the renowned Dutch tulips, correct? Throughout my stay, I endeavoured to maximize my experience by exploring the famed cities, all easily accessible and budget-friendly by train. I even had the opportunity to venture to Brussels over the weekend, a mere 90-minute train ride away! During my time in the Netherlands, I had the privilege of attending the celebrated Dutch Flower Parade, known as "Bloem
Read more

POEMA Final Workshop

Image
 From September 5th to 9th, 2022, the final POEMA workshop was held in the romantic city of Paris😍. The Sorbonne University Pierre and Marie Curie Campus provided a fitting location for this open international conference on Polynomial Optimization, which aimed to disseminate the network's results. Almost all ESRs planned to present their final results, making it exciting and significant to see everyone from the project in person again. The daily presentations were informative, providing opportunities to connect with other researchers and expand our network. We had many friendly conversations with the other ESRs, discussing our challenges and life😊. On September 7th, I was pleased to present " A Primal-Dual Augmented Lagrangian Method for the solution of Convex Semidefinite Programming Problems ." The feedback and questions I received were invaluable, allowing me to delve deeper into certain parts of my results. ( The YouTube link of my presentation ). On September 8th,
Read more