论文部分内容阅读
量子粒子群算法在优化过程中需要权衡局部探索性和全局开拓性,进化后期由于全局开拓能力的丧失使得种群多样性减少,设计了一种基于欧式距离的混合量子粒子群算法,通过计算粒子的种群多样性,当种群多样性低于阈值范围时加入基于欧式距离的种群划分策略划分子种群,从而保证获得全局最优解。利用标准测试函数验证提出的混合量子群算法有效性。提出了基于混合量子粒子群的Mean Shift算法(HQPSO Mean Shift)完成目标快速跟踪,克服传统Mean Shift算法的在跟踪快速移动目标时出现“跟丢”的问题。
Quantum particle swarm optimization algorithm needs to balance the local exploratory and global pioneering. At the late stage of evolution, the population diversity is reduced due to the loss of global exploiting ability. A hybrid quantum particle swarm optimization algorithm based on Euclidean distance is designed. When the diversity of the population is below the threshold, a population-based strategy based on the Euclidean distance is added to divide the subpopulation so as to ensure the global optimal solution. The validity of the proposed hybrid quantum group algorithm is verified by using standard test functions. This paper proposes a Mean Shift algorithm (HQPSO Mean Shift) based on hybrid quantum particle swarm optimization (PSO) to complete the target fast tracking and overcomes the problem of “missing following” when tracking the fast moving target by the traditional Mean Shift algorithm.