When particle swarm optimization algorithms are employed in robots to search and track moving targets,the algorithms cease to converge during later iterations.To alleviate this problem,this paper proposed a search algorithm that integrates the Newton method into the traditional particle swarm optimization algorithm.This algorithm utilized a Markov chain that allowed the robots to randomly select in each iteration,with a given probability,the Newton method or the traditional particle swarm optimization algorithm to calculate their subsequent positions.To emulate a realistic environment that robots might need to operate in,the algorithm introduced a communication term to maintain a connectivity between robots and their controlling station allowing the relay of target information.Experimental simulations show that the improved algorithm can more effectively search and track moving targets whilst alleviating the aforementioned problem.%粒子群算法引导机器人搜索跟踪动态目标时,在迭代后期易出现收敛停滞现象.为了改善上述情况,提出了结合牛顿法的改进粒子群算法.为了结合粒子群算法与牛顿法,在算法中引入了马尔可夫链,这使得机器人在每一次迭代时以一定的概率随机选择牛顿法或粒子群算法搜索跟踪目标.为了模拟机器人搜索动态目标的真实环境,还利用了通信项使机器人以一定的方式努力与基站保持通信,用来实时更新目标信息.仿真结果表明,改进的粒子群算法能有效地寻找并跟踪动态目标.
展开▼
