BEGIN:VCALENDAR VERSION:2.0 PRODID:-//sebbo.net//ical-generator//EN URL:https://lectures.london/imperial-college/quantum-optimization-benchmar king-library-the-intractable-decathlon/calender.ics NAME:Lectures London X-WR-CALNAME:Lectures London TIMEZONE-ID:Europe/London X-WR-TIMEZONE:Europe/London BEGIN:VEVENT UID:0008ec65-778f-49d3-9344-c62f88ca43bc SEQUENCE:0 DTSTAMP:20260611T073601 DTSTART:20260615T150000 DTEND:20260615T160000 SUMMARY:Quantum Optimization Benchmarking Library – The Intractable Deca thlon LOCATION:Imperial College: LT3\, ACE Extension DESCRIPTION:Title: Quantum Optimization Benchmarking Library – The Intra ctable Decathlon\nAbstract: Optimization is the methodological backbone be hind the operations of major industrial areas\, including Logistics\, Manu facturing\, Power Systems\, and Process Systems Engineering (PSE) [1]\, am ong others. In the face of numerous challenging Combinatorial Optimization problems\, the development of novel\, specialized hardware has been an ac tive research area over the past decade\, leading to fundamental advanceme nts in Quantum Computing and other Physics-inspired devices and algorithms [2]. Optimization is a well-established domain in quantum applications re search\, particularly for existing and near-term quantum hardware [3]. The field is driven by heuristic quantum algorithms compatible with the hardw are\, such as the Quantum Approximate Optimization Algorithm\, or QAOA [4] . Recent hardware improvements pave the way for thorough benchmarking of h euristic quantum algorithms at scale [5].\nThis work presents ten optimiza tion problem classes that are difficult for existing classical algorithms and can be linked to practically relevant applications\, aiming to enable a fair\, comparable\, and meaningful benchmarking effort for quantum optim ization methods. The initial compilation contains instances of the problem s known as Market Split\, Low Autocorrelation Binary Sequences\, Minimum B irkhoff Decomposition\, Steiner Tree Packing\, Sports Tournament Schedulin g\, Portfolio Optimization\, Independent Set\, Network Design\, Vehicle Ro uting Problem\, and Topology Design. These are primarily typical Combinato rial Optimization problems best known through their Mixed-Integer Programm ing (MIP) formulations\, which were then cast into a corresponding QUBO fo rmulation [6].\nWhile these problem classes vary in their individual prope rties\, such as objective and variable types\, coefficient ranges\, and de nsity\, they all become challenging for established classical methods at s ystem sizes of approximately 100 to 10\,000 decision variables. The small sizes at which difficult problem instances appear enable testing quantum a lgorithms for these problems already today. This is not\, however\, a stat ic collection\, and the archive is expected to grow over time with contrib utions. We envision the OR community being able to provide this database w ith interesting and relevant instances and solutions to the challenging pr oblems within it\, as it did\, for example\, in the composition of the lib rary of Mixed-Integer and Continuous Nonlinear Programming Instances (MINL PLib) [7].\nIn this talk\, we reference the results from state-of-the-art solvers for instances from all problem classes and demonstrate exemplary b aseline results obtained with quantum solvers for selected classes. The ba seline results illustrate our suggested benchmark reporting\, aiming for c omparability of the used methods\, reproducibility of the respective solut ions\, and trackability of algorithmic and hardware improvements. We encou rage the optimization community to explore the performance evaluation of a vailable classical or quantum algorithms and hardware with the benchmarkin g problem instances presented in this library.\nThis work was an effort of the IBM Quantum Optimization Working Group and resulted in the paper “Q uantum Optimization Benchmark Library The Intractable Decathlon\,” which has already been submitted to ArXiV but is still pending publication\, as well as the Quantum Optimization Benchmark Library (QOBLIB) repository [8 ].\n More info\n \n \n \n \n \n Refer ences\n \n \n \n \n Speaker bio\n \n \n \n\n \n \n References\n \n \n \n [1] https://aiche.onlinelibrary.wiley.com/doi/full/10.1002 /aic.17651\n[2] https://arxiv.org/abs/2310.03011\n[3] https://www.nature.c om/articles/s42254-024-00770-9\n[4] https://arxiv.org/abs/1709.03489\n[5] https://arxiv.org/pdf/2502.06471\n[6] https://arxiv.org/abs/2307.02577\n[7 ] https://www.minlplib.org/index.html\n[8] https://git.zib.de/qopt/qoblib- quantum-optimization-benchmarking-library\n\n \n \n \n \n Speaker bio\n \n \n \n David E. Bernal Neira is an Assistant Professor in the Davidson School of Chemic al Engineering at Purdue University. His research centers on mathematical optimization\, artificial intelligence\, and computational methods for sol ving scientific and engineering problems\, with applications in process sy stems\, energy\, and chemical engineering. His core expertise is in nonlin ear discrete optimization\, encompassing theory\, algorithms\, and softwar e. He also leads research in quantum computing\, with emphasis on quantum algorithms for optimization\, computational chemistry\, and machine learni ng. He has co-authored peer-reviewed publications\, developed open-source tools\, and delivered invited talks across academia\, government\, and ind ustry. He has taught several courses\, including one he co-designed on opt imization\, quantum computing\, and machine learning. He collaborates broa dly with researchers in academia\, national laboratories\, government agen cies\, and industry. At Purdue\, he leads the SECQUOIA lab (Systems Engine ering via Classical and QUantum Optimization for Industrial Applications). URL;VALUE=URI:https://www.imperial.ac.uk/events/209913/quantum-optimizatio n-benchmarking-library-the-intractable-decathlon/ END:VEVENT END:VCALENDAR