Using Reinforcement Learning to Perform Qubit Routing in Quantum Compilers

‘‘Qubit routing” refers to the task of modifying quantum circuits so that they satisfy connectivity constraints a target computer. This involves inserting SWAP gates into circuit logical only ever occur between adjacent physical qubits. The goal is minimise depth added by gates. In this article, we propose qubit routing procedure uses modified version deep Q-learning paradigm. system able outperform procedures from two most advanced compilers currently available (Qiskit and t \( | \) ket \rangle ), on both random realistic circuits, across range near-term architecture sizes (with up 50 qubits).