基于小生境粒子群技术的多航迹规划研究

Multiple Routes Planning for Air Vehicles Based on Niche Particle Swarm Optimization

文章将飞行器多航迹规划转化为多峰值函数优化问题,并以此为基础提出基于小生境粒子群技术的多航迹规划方法。该方法采用特定的粒子编码方式和适当的适应度函数,在满足各种航迹约束的条件下,通过引入RCS(Restricted Competition Selection)小生境生成策略,将航迹规划空间内的粒子群形成不同的相互独立的小生境子种群。在进化过程中,所有粒子个体只在各自的小生境子种群内部进化,追逐不同的极值点。当进化结束时,每个小生境子种群将分别生成一条各自的最优航迹,从而为飞行器生成了多条不同的可选航迹。仿真结果表明了该方法的有效性。

Air vehicle routes planning can be seen as a multiple-peak function optimization problem. The existing genetic algorithm for this problem is complicated. We present a new and more simple algorithm for multiple routes planning of air vehicles which is based on niche particle swarm optimization. In this algorithm, specific particle swarm coding representation and suitable fitness function was used. It can process each kind of the mission con- straints. By introducing the Restricted Competition Selection （RCS） niche technique, the individuals of the population form a number of sub-populations. In the evolutionary process, all routes evolve only in their own sub-population. At the end of the evolution, each sub-population provides an optimal route, and thus the algorithm generates multiple routes for the vehicle. The simulation results demonstrated the feasibility of the algorithm.