QISA: Incorporating quantum computation into Simulated Annealing for optimization problems

This paper proposes a novel model of QISA (Quantum-Inspired Simulated Annealing) which embeds quantum computation into the Simulated Annealing (SA) process for optimization problems. Compared with previous SA studies, QISA adopts Quantum Bits (Qubits) rather than the conventional binary bits as the coding scheme, and update the entries in the Qubits sequence with different probabilities by the proposed methods of heat observation and quantum rotation gate. By flexibly controlling the heating function and rotation angle, QISA does not have to either accept or reject, but can partially accept a new solution. The careful experiments on a number of numerical optimization and combinatorial optimization problems validate its better convergent performance and lower sensitivity to parameters.