分布式传感器调度模型与自适应概率粒子群优化算法

不同体制的多个传感器通常部署于不同位置,据此采用分布式计算思想研究其调度问题.设计了传感器指控模块和传感器模块,探讨两者间的信息交互过程,给出基于最小调度时间间隔的传感器探测任务分解方法,建立传感器探测目标的匹配度计算模型.针对调度方案生成子模块设计了一种自适应概率粒子群算法,算法中粒子的分量根据方案适应值大小以不同的概率取相应值,体现粒子在迭代过程中的思考.实例分析表明,该算法能在迭代前期较快地收敛到一个较优值,这一特点使得在迭代次数有限的情况下,算法仍可获得较好的调度方案,满足调度方案实时高效的要求.

A Model of Distributed Sensors’Scheduling and Self-Adaptive Probability Particle Swarm Optimization Algorithm

This paper introduced a distributed computing method to study the scheduling problem in multi-sensor systems.According to the features of these sensors which were usually deployed at different locations,it redesigned C2 (Command and Control) module and sensor module,and discussed information interaction procedure between two modules.Then it established a sensor-target detection match degree computing model which involved task decomposition and minimal scheduling period methods.A self-adaptive probability particle swarm optimization (SAPPSO) algorithm for scheduling program was given.In SAPPSO,particle fitness value was based on different probabilities,which reflected the thinking of particle during iteration process.Experimental results showed that SAPPSO algorithm converged quickly,especially in the previous iteration period,which enabled SAPPSO to fulfill the requirements of real-time and high efficiency for scheduling.