Optimizing Quantum Circuit Placement via Machine Learning
TimeTuesday, July 12th11:37am - 12pm PDT
Location3005, Level 3
DescriptionBy formulating QCP as a bilevel optimization problem, this paper proposes a novel two-step machine learning (ML)-based framework to tackle this challenge.
we adopted a policy-based deep reinforcement learning (DRL) algorithm to optimize the SWAP strategy.
A genetic algorithm is then deployed to solve the upper-level search problem,
which optimizes the initial mapping with a lower SWAP cost.
Compared with the heuristic approaches, our ML-based method significantly reduces the SWAP cost.
In comparison with the exact search,
our proposed algorithm is able to achieve the same optimality while reducing the runtime cost by up to 30 times.