Quantum-Inspired Artificial Lemming Optimization for Topology-Constrained Quantum Circuit Mapping
Xuena Han, Pengxiang Han, Hui Li, Jinchen Shang, Hewen Bai · IEEE Access · 2026
This paper studies the qubit-mapping problem under restricted hardware connectivity and proposes a quantum-inspired artificial lemming optimization method, denoted as QIALO. The term quantum-inspired is used to indicate that the algorithm is executed on a classical computer while borrowing representation and search mechanisms from quantum information, including Bloch-sphere-based individual encoding and quantum-rotation-based mutation. To make the optimizer applicable to discrete qubit mapping, this paper explicitly defines the continuous-to-discrete decoding procedure, the topology-constrained routing process, and the routing-oriented fitness function. Within the ALO framework, QIALO introduces Bloch spherical-coordinate encoding, a $t$ -distribution-based quantum-rotation mutation strategy, an adaptive search-direction factor, and dynamic foraging-radius adjustment. In addition to the original benchmark comparison, the experimental evaluation includes repeated-run Mean±Std statistics, a group-level SABRE/depth comparison, and a component-level ablation analysis. The grouped statistics show that QIALO reduces the average SWAP overhead by 17.0% compared with ALO over all 30 benchmarks and reduces the Std/Mean ratio from 8.4% to 5.7%. The ablation results further show that the full QIALO obtains the lowest SWAP, CNOT, and depth values among the tested variants, and CEC2017 convergence-curve visualizations are included to illustrate the general search behavior of the optimizer. These results suggest that QIALO can improve routing quality and stability on the evaluated topology-constrained mapping cases, while broader validation on additional hardware-aware settings remains necessary.