多智能体蝙蝠算法在无线传感器中的应用

The Multi-Agent Bat Algorithm Applied to Wireless Sensor Networks

针对无线传感器网络(WSNs)节点的定位误差较大的问题,提出了一种新的具有局部搜索能力强的多智能体蝙蝠算法。改进算法中对寻优蝙蝠个体融入多智能体技术,通过邻域竞争合作算子以及自学习过程提高了算法全局搜索能力,避免算法陷入局部最优,加快算法的收敛速度。通过对标准测试函数的仿真,改进算法相比于其他算法,寻优精度和进化效率得到了较大的提高。随后采用多智能体蝙蝠算法求解无线传感节点定位问题,仿真结果表明改进算法减少了测距误差对定位精度的影响,提高了未知节点定位的精度,为无线传感网络节点定位的实际应用提供理论参考。

In order to solve the node location error in wireless sensor networks （WSNs）, this paper proposes a new multi-agent bat algorithm, which possesses favorable local searching ability. In the proposed algorithm, the bat indi- vidual is a agent, which could compete and cooperate with its agent neighbor areas to improve the efficiency of local searching. In this way the multi-agent bat algorithm （MA-BA）could avoid the algorithm into a local optimum and in- crease the convergence speed. Simulation results for standard test functions indicate that the proposed algorithm re- markably improves the global optimizing ability and evolutionary efficiency compared to other algorithms. Through implementing the MA-BA to node location prediction, the precision of the unknown node location could be im- proved due to decreasing the ranging error and has a certain significance to practical application of wireless sensor network node localization.