多波束粒子群深度零点轨迹信息搜索效率优化

Optimization of Multi Beam Particle Swarm Depth Zero Trajectory Information Search Efficiency

传统的基于粒子群算法的前馈神经网络训练系统进行数据库访问时,易陷入局部极值,产生零点轨迹信息搜索效率较低,局部极小点和搜索方向紊乱。提出一种改进的粒子群优化算法。构建基于误差反传的神经网络系统结构,引入混沌映射概念,提出了一种根据粒子搜索状态,动态调整粒子飞行速度和位置的粒子群优化算法,提高多波束粒子群深度零点轨迹信息的提取的搜索效率,根据粒子的轨迹信息,研究如何动态调整粒子的搜索速度和方向,提高了训练和控制精度与效益。仿真实验表明,该算法进行多波束粒子零点轨迹信息搜索,效率较高,通过外力的干涉尝试调整粒子的方向,使得粒子可以逃离这个稳定阶段,提高了粒子收敛速度,提高控制搜索精度,运行时间较短。算法在智能控制等领域具有较好的应用价值。

The traditional database access in particle swarm optimization algorithm based on the feed forward neural net?work training system, easy to fall into local extremum, produce zero trajectory information search efficiency is low, local minima and the search direction of disorder. This paper proposes an improved particle swarm optimization algorithm. The construction of the error back propagation neural network based on system structure, introduces chaos mapping concept, proposes a search condition according to the particle, particle swarm optimization algorithm to dynamically adjust the parti?cle flight speed and position, improve the search efficiency of multi beam particle swarm to extract depth zero trajectory in?formation of the particle trajectory, according to the information, to study how the dynamic adjustment of the search the speed and direction of the particle, improve the training and the control precision and efficiency. Simulation results show that the algorithm for multi beam particle zero trajectory information search, it has high efficiency, through the external in?terference try adjusting the particle direction, so the particles can escape from the stable stage, particle convergence speed is increased, the control of search accuracy is improved, the running time is short. The algorithm has good application value in the fields of intelligent control.