基于多智能体遗传算法的无人机突防航线规划  

UAV Low- altitude Penetration Route Planning Based on Muilt- Agent Genetic Algorithm

作　　者：饶卫平[1] 杨任农[1] 雷晓义[1] 张滢[1] 

机构地区：[1]空军工程大学航空航天工程学院,西安陕西710038

出　　处：《计算机仿真》2015年第4期39-43,共5页Computer Simulation

摘　　要：针对无人机突防航线规划中需要威胁规避以及低空突防的问题,结合智能体自学习自适应的特点,提出了基于多智能体遗传算法来解决航线规划问题。采用自适应交叉和变异,改进自学习算子中获取下一代智能体的方法,将智能体和遗传算法的优点结合起来,达到了全局最优的效果。通过和传统遗传算法进行仿真比较,相比之下,基于多智能体的遗传算法可以有效规避威胁源,并能实现低空突防的效果,整体所需的航程代价小,运算速度快。In order to satisfy the demand of threat avoidance and low - altitude penetration in unmanned aerial vehicle（UAV） route planning, a route planning optimization method is proposed based on multi -agent genetic algorithm with full consideration of the self - learning and self - adapt feature of the muhi - agent. By adopting self - adapt crossing and mutation operator, the nextgeneration agent is acquired by improved self- learning operator, which combines agent with genetic algorithm to achieve global optimal route. Simulation results show that the im- proved genetic algorithm can effectively avoid threat and fulfill low - altitude penetration tasks with lower cost and less operation time compared with traditional ones.

关 键 词：航线规划 多智能体 遗传算法 低空突防 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]