基于量子粒子群的信赖域算法

Trust region algorithms based on quantum-behaved particle swarm optimization

通过量子行为能增强粒子的全局寻优能力,引进了量子粒子群算法(QPSO),用于求解信赖域(TR)算法的子问题,并将这2种算法有效结合.数值实验表明,新算法具有良好的全局寻优能力,并有效提高收敛速度和避免早熟.

Through acts of quantum particles can enhance the ability of global optimization, introdueted a quantum particle swarm optimization （ QPSO ） for solving trust region （ TR ） algorithm for a subset of issues, and combined the two techniques. Then with the proper setting, the experimental results indicated that the new algorithm has good convergence and fast convergence rate.