IEEE Congress on Evolutionary Computation (CEC)
QiSA: A Parameter-free Quantum-inspired Search Algorithm – A Preliminary Study on Combinatorial Optimization
Kuo-Chun Tseng, Wei-Chieh Lai, Chen-Hsin Lu, Wei-Chun Huang, Jen-Shin Hong, Hsin‐Hung Cho
Abstract
This study is inspired by quantum algorithms and proposes the Quantum-inspired Search Algorithm (QiSA), an entropy-driven quantum-inspired metaheuristic for parameter-free combinatorial optimization. Guided by Occam’s razor, QiSA simplifies the search process and reduces parameter dependence as much as possible in this preliminary study. It uses entropy as the core signal to unify parameter control, overcoming the traditional dilemma of balancing exploration and exploitation, and maps it to a confidence-like level to define an interpretable closed-loop termination criterion. QiSA also exhibits several observable quantum-like behaviors during the search, reflecting characteristic phenomena associated with quantum algorithms. Experiments on classic combinatorial optimization benchmarks show that QiSA delivers robust, consistent, and competitive performance, with the potential to stand alongside Differential Evolution in combinatorial optimization.