There are numerous optimization issues in finance, logistics, biotechnology, and AI the place you must discover the most effective mixture from an unlimited vary of decisions. Combinatorial optimization issues akin to these are troublesome to resolve at excessive velocity and at an inexpensive computational price with current computer systems as a result of the variety of combinatorial patterns will increase exponentially as the dimensions of the issue grows.
One option to sort out these combinatorial optimization issues is to map them to a binary illustration referred to as an Ising mannequin, after which use a specialised optimizer to seek out the bottom state of this Ising system.
Toshiba’s new Simulated Quantum Bifurcation Machine+ (SQBM+) on Azure Quantum, based mostly on its Simulated Bifurcation Machine (SBM), is an Ising mannequin solver that may clear up advanced and large-scale combinatorial optimization issues with as much as 100,000 variables at excessive velocity.
Toshiba has adopted a brand new strategy, impressed by their quantum computing analysis, that considerably improves the velocity, accuracy, and scale of their SBM. There are two algorithms out there by the SQBM+ supplier in Azure Quantum: the high-speed Ballistic Simulated Bifurcation algorithm (bSB) designed to discover a good answer in a short while; and the high-accuracy Discrete Simulated Bifurcation algorithm (dSB) which finds extra correct options at a calculation velocity that surpasses that of different machines (each classical and quantum). An auto-tune perform has additionally been applied that may auto-select which algorithm to make use of based mostly on the issue submitted. These algorithms are optimized routinely to supply the most effective efficiency on GPU {hardware} deployed within the Azure cloud.
Customers can choose one among these algorithms particularly, or just enable the auto choice perform to decide on on their behalf. This selection is made by supplying values for the “algo” and “auto” parameters throughout solver instantiation utilizing the Azure Quantum Python SDK. Extra info is on the market within the Toshiba SQBM+ supplier documentation, and a pattern displaying how to decide on between the totally different algorithm choices may be discovered on the qio-samples repo.
“The core expertise of SQBM+ is SBM, which is software program that makes use of at present out there computer systems and achieves high-accuracy approximate options for advanced and large-scale issues in a brief period of time. The result is the flexibility to resolve Ising issues of as much as 100,000 variables—at roughly a 10X enchancment over our current PoC service. And that is now all simply accessed by the Azure Quantum cloud platform,“—Shunsuke Okada, Company Senior Vice President and Chief Digital Officer of Toshiba.
Azure Quantum clients can entry SQBM+ by including the supplier to their Quantum Workspace and deciding on one of many out there pricing plans: “Study & Develop” (experimentation) and “Efficiency at scale” (business use).
Since becoming a member of the Azure Quantum Community in September 2020, Toshiba has constantly improved its quantum-inspired optimization solvers expertise. Prospects who wish to clear up combinatorial optimization issues together with dynamic portfolio and danger administration, molecular design, and optimizing routing, partitioning, and scheduling in a variety of fields can apply SQBM+ as we speak, harnessing the GPU sources within the Azure cloud by Azure Quantum.
Study extra and get began as we speak with Toshiba’s SQBM+ on Azure Quantum.