Search papers, labs, and topics across Lattice.
Distributed quantum computing (DQC) integrates the computational resources of multiple small-to-medium-scale quantum processors to form a larger composite system capable of solving problems beyond the capacity of a single device. However, remote quantum communication between processors incurs significant overhead and is highly sensitive to entanglement degradation and link noise. Existing compiler-level qubit allocation strategies typically overlook these fidelity variations, leading to inefficient or unreliable mappings. In this work, we propose a Fidelity-aware Qubit Allocation System (FQAS). This system explicitly models how purification requirements and resource consumption affect the cost of remote communication in DQC, and it is the first to integrate a fidelity model into the distributed qubit allocation. FQAS further incorporates a Fidelity-aware Path Planning (FPP) algorithm that identifies communication paths with minimal resource consumption. The resulting Fidelity-aware Qubit Allocation Problem (FQAP) is formulated as a generalized quadratic assignment problem and solved using a customized hybrid genetic algorithm (HGA). We evaluate FQAS across various quantum circuits and network topologies. Experimental results demonstrate that the proposed system effectively reduces remote communication overhead, achieving an average improvement of approximately 12.5% on real circuits and 5.3% on random circuits compared with existing methods.
