基于Lévy飞行特征的蝙蝠算法及其在WSN定位中的应用∗

摘要

Given the shortcomings of premature convergence and slow convergence speed in bat algorithm, an im-proved adaptive bat algorithm( BA) based on Lévy flight strategy characterized by heavy-tailed distribution is pro-posed,which differs traditional BA in update approach of bat’s flying velocity and positions. It could effectively keep from the algorithm into a local optimum and accelerate convergence to achieve a balance between exploration and exploitation mechanisms. In WSN applications,we converted the WSN location problems into the global optimization ones and by applying ZigBee hardware platform to compare with other algorithms in different position,a conclusion could be drawn that the improved algorithm is of quicker convergence speed and higher precision,moreover,which realizes simple condition,high accuracy with huge value of practical engineering applications.%针对蝙蝠算法收敛易早熟、收敛速度慢等不足，提出一种改进的基于Lévy飞行特征自适应的蝙蝠算法。采用Lévy飞行策略取代原算法中蝙蝠飞行速度和位置的更新方式，充分利用Lévy飞行的重尾效应，有效避免局部最优值的吸引，加快了收敛速度，达到寻优能力和搜索能力的平衡。在无线传感器网络自身定位应用中，把定位问题转换为一个全局优化问题，使用改进的算法进行定位计算。通过Zigbee平台的实验表明，改进后的算法在不同空间位置的定位精度更高，收敛速度更快。算法实现条件简单、精度高，具有较高的实际工程应用价值。