Optimally Stopped Optimization
Walter Vinci, Daniel A. Lidar · Physical Review Applied · 2016
Quantum computing may be the only pragmatic way to solve some problems, but when it is not absolutely necessary, is it actually worthwhile? The authors integrate the fields of heuristic optimization and optimal stopping to build a general framework for benchmarking randomized optimization algorithms. Their approach avoids bias and arbitrariness, and is particularly suited to determining the break-even point at which quantum optimization is superior to classical, when both raw performance and technology costs are taken into account.