Abstract: In order to overcome the limit that the traditional quantum-inspired evolutionary algorithms (QEA) only use the current best individuals to guide the evolution, which leads to the insufficient evolutionary capability and may often end up by providing sub-optimal solutions, this paper proposes a self-adaptive QEA combining the strategy of keeping away from the worst (KSQEA), which evolves individuals not only close to the best but also far away from the worst. In this way, KSQEA is able to acquire the evolutionary driving force even if the advantage of the best individuals is not obvious. In addition, this paper proposes a wavy rotation angle decaying method to balance the exploration and the exploitation of search. The experimental results both on the function optimization and the knapsack problem show that these measures successfully increase the searching capability of the algorithm and the quality of solutions.

Key words: evolutionary algorithm, quantum-inspired evolutionary algorithm, self-adaptation, function optimization, knapsack problem

摘要： 针对传统的量子进化算法只使用当前最优个体作为指导，存在进化能力不足，易陷入局部极值的问题，提出了一种结合远离最差策略的自适应量子进化算法KSQEA，使个体在进化过程中不仅向最优个体靠近，而且还远离最差个体，这样在最优个体优势不明显时仍有可能获得进化动力。旋转角更新则采用一种新的自适应波浪式衰减方式，以更好地平衡探查和利用。在函数优化和背包问题上的实验结果表明，以上措施有效地增强了算法的搜索能力，提高了解的质量。

关键词: 进化算法, 量子进化算法, 自适应, 函数优化, 背包问题