量子行为引力搜索算法

Quantum-behaved gravitational search algorithm

为提高引力搜索算法的优化能力,通过在原始算法中融合量子计算,提出一种量子行为引力搜索算法.该算法采用类似量子行为粒子群优化的寻优机制,在每步迭代中,计算个体适应度,根据适应度计算个体质量,取前K个质量最大的个体作为候选集.采用轮盘赌方法在候选集中选择一个作为Delta势阱的中心,调整其他个体向该中心移动完成一步优化,在优化过程中使K值单调下降,以期达到探索和开发的平衡.标准函数极值优化的实验表明,所提出的算法比原算法在优化能力和优化效率两方面都有明显提高.

To enhance the optimization performance of the gravitational search algorithm, by introducing quantum computing to the original algorithm, a quantum-behaved gravitational search algorithm is proposed. The optimization strategies of the proposed algorithm are similar to the quantum-behaved particle swarm. In each of iteration, the fitness of each individual is evaluated, and then the quality of each individual is calculated based on its fitness. The first K greatest individuals are taken as a candidate set, in which an individual is randomly selected with roulette and is taken as the center of the Delta potential well. By adjusting other individuals move to the potential well centre, a single-step optimization is completed. In the process of optimization, the value of K is made decreasing monotonically to achieve a balance of exploration and exploitation. The experimental results on benchmark functions extreme optimization show that the proposed algorithm is obviously superior to the original one in performance and efficiency.