Abstract: A self-adaptive variable step double chains quantum genetic algorithm is proposed in this paper, which improves the slow convergence rate and the poor robustness of the double chains quantum genetic algorithm based on real-code and gradient information. In the algorithm, the mathematical model is constructed which reflects change rate of objective fitness function. Coefficient k of variable step is established to reflect the relative change rate of the fitness at the current searching place. The searching process of the optimal solution can be improved by adjusting the coefficient k of variable step which affects the relative change rates of the fitness. The updating strategies of quantum revolving gate [Δθ] are constructed. The specific procedure of the algorithm is designed for the complex continuous space optimization problems. The results of simulation show that the algorithm can improve the robustness effectively and improve the convergence rate.

Key words: quantum computation, double chains quantum genetic algorithm, variable step, optimization algorithm

摘要： 为了克服基于实数编码和目标函数梯度信息的双链量子遗传算法存在收敛速度慢和鲁棒性较差的缺点，提出了一种自适应变步长双链量子遗传算法。建立了反映目标适应度函数变化率的数学模型；构造了反映当前搜索点处适应度相对变化率的变步长系数k，通过调整k以改善适应度函数相对变化率从而优化解的搜索过程；提出了在迭代过程中的量子旋转门转角[Δθ]更新策略。针对复杂连续函数的优化问题，设计了算法的具体实施步骤，并对典型复杂函数进行了仿真。结果表明，该算法有效地改善了双链量子遗传算法的鲁棒性，加快了算法收敛速度。

关键词: 量子计算, 双链量子遗传算法, 变步长, 优化计算