Annotation

Annotation: This paper proposes a real‑valued single‑objective optimization algorithm that combines the potential of quantum‑inspired computation with a context‑aware self‑tuning mechanism. The core of the new algorithm is a modified quantum‑inspired genetic algorithm employing multi‑level quantum systems and a physically grounded decoherence model that emulates the noisy intermediate‑scale quantum era. In contrast to existing quantum‑inspired optimization algorithms that rely on manual calibration or fixed heuristics, the proposed algorithm automatically adjusts key control parameters, such as the rotation angle in quantum gates and the mutation probability, by analyzing its own performance history using success‑history adaptation combined with a second‑order Lehmer weighted mean. This enables dynamic balancing between global exploration and local exploitation tailored to the characteristics of the objective function landscape. Comprehensive evaluation on a suite of benchmark functions from the evolutionary computation benchmark set demonstrates that the proposed algorithm achieves high reliability and robustness on multimodal, as well as complex hybrid and composite functions. The results highlight the promise of integrating quantum‑inspired optimization models with adaptive control strategies to develop robust black‑box optimization tools.