We run continuous-time quantum walks with perfect and peak state transfer on a present-day superconducting quantum processor, IBM Quantum's
156-qubit Heron device ibm_kingston, deliberately without error correction or error mitigation. Perfect state transfer and its relaxations have been studied extensively in algebraic graph theory, where one can prove exactly which graphs move a quantum state from one vertex to another with probability one (or with the largest probability the graph allows, in the case of peak state transfer), and at what time. Our experiments cover perfect state transfer on the hypercubes
Q2=C4 and
Q3, and on the Hamming-weighted paths
HPn, the equitable-partition quotients of the hypercubes
Qn, for
n up to
20; peak state transfer on the
Xn family for
n up to
11; and, for comparison, moving the state along a path of qubits by a sequence of SWAP gates.
Abstract In this paper, within the framework of continuous-time quantum walks (CTQWs), we decompose graphs into disjoint strata by adopting the idea of equitable partitions. Based on the thought, we investigate the success probability of finding marked vertices on the N -fold star power graph G ⋆ N (where G is a cycle graph with three vertices), as well as the time required for perfect state transfer. Additionally, we explore the time for perfect state transfer on G ⋆ N under partition conditions from the perspective of quantum probability. Specifically, we surprisingly discover an interesting result: the times for perfect state transfer under quantum probability-based partitions and equitable partitions are completely consistent when the coupling strength J = 1 and the parameter λ = 1.