D-Wave upgrades Stride hybrid quantum solver, claims 200x gain
|

D-Wave upgrades Stride hybrid quantum solver, claims 200x gain

Assembly line balancing: from 120 seconds to 15 seconds

Three-course learning path opened for applied quantum optimization

D-Wave Quantum Inc. said on Nov. 4 (local time) that it has improved the core algorithms of its Stride hybrid solver, optimization software that combines quantum and classical computing. The company said the improved version found better solutions in 15 seconds than the previous version had produced in 120 seconds.

According to the company, Stride is a hybrid solver that uses quantum annealing hardware (quantum computing that works by seeking the lowest-energy state) together with classical algorithms. It is available on Leap, D-Wave’s quantum cloud service. On the traveling salesman problem (TSP, finding the shortest route that visits each of a set of points once and returns to the start) with up to 10,000 locations, the improved version produced solutions within 0.02% of the best solution on average in 15 seconds. The company said that, for the same amount of time, this was more than 200 times closer to the best solution than the previous version. For problems with more than 10,000 locations, it came within 3.8% on average in 120 seconds, more than 500 times closer than the previous version. In assembly line balancing problems, too, its 15-second solutions for simple cases were better than the previous version’s 120-second solutions.

The company also launched a three-course learning path called “Applied Quantum Optimization.” It covers finding use cases, mathematical modeling and building apps with Stride, and some courses include Leap access and office hours (Q&A sessions). Alex Condello, D-Wave’s senior vice president of algorithms, performance and tools, said in effect that developers need proven technology that delivers better answers faster.

Optimization problems in fields such as logistics and manufacturing are where the quantum computing industry is looking first for commercial use cases. However, these figures describe performance “compared with the previous version” as presented by the company. No comparison with classical optimization software or independent verification has been confirmed, and publicly available materials do not show how much the quantum hardware and the classical algorithms each contribute within the hybrid solver. The key question is whether external benchmarks emerge.

Related articles

사이언스모니터 뉴스레터
AI·양자·R&D 핵심 해설을 주 2회(화·금) 이메일로 받아 보세요.

Similar Posts