基于MAS理论与动态联盟的传感器自适应任务调度算法

Sensor Adaptive Task Scheduling Algorithm Based on Multi-agent System Theory and Dynamic Alliance

针对无线传感器网络任务调度的实效性及节点能量有限的特点,通过多代理系统(MAS)进行任务划分与逐层处理,根据动态粒子群的自适应优化理论,提出一种传感器自适应任务调度算法。该算法基于多代理的网络架构,根据动态联盟的数学模型,将离散粒子群算法的自适应性与动态联盟的应变能力相结合,通过适应值函数及粒子的更新方法获得全局搜索,实现任务的动态最佳自适应分配。实验结果表明,该算法在降低任务的总执行时间、节点负载压力及网络的总能量消耗量上取得较好的效果。

For the characteristics of the effect of task scheduling and node energy limited in Wireless Sensor Network（WSN）,through the Multi-agent System（MAS）for the division of tasks and processing step by step,and the dynamic theory of the adaptive particle swarm optimization,this paper proposes a sensor adaptive task scheduling algorithm.The algorithm is based on MAS architecture,through mathematical models of dynamic alliance,adaptive discrete particle swarm algorithm is combined with dynamic alliance to obtain better global search method by adapting to update the value of the function and particle achieve optimal adaptive dynamic task allocation.Experimental results show that this algorithm achieve better effects in task operation time reduction,node load pressure reduction and energy consume of networks.