【作者】 罗泉；

【导师】 刘忠；

【作者基本信息】 国防科学技术大学 ， 管理科学与工程， 2010， 硕士

【摘要】 空对地攻击能力是空中力量由积极防御型向攻防兼备型转化的决定性因素,受到世界各国的高度重视,研究空对地攻击行动的相关技术对提高我军战斗力具有重大意义。本文从作战计划筹划的角度考虑如何有效运用航空兵作战力量实施空对地攻击任务。对地攻击行动的规划,包括航迹规划和任务分配两个方面,两者既互相耦合且都是十分复杂的问题,本文又特别考虑战场空间内存在多种威胁的实际情况,这使航迹规划和任务分配的难度显著增加。本文提出了一种综合飞行器航迹规划与任务分配问题的联合求解方法:先使用基于动态规划的航迹规划方法规划出任意两个目标之间的航线,然后使用基于多子群蚁群算法的任务分配方法为飞行器分配任务。这样可以将飞行器性能约束、战场环境约束等约束条件在航迹规划过程中进行处理,而在任务分配过程中集中处理与任务相关的约束以及编队协同作战等问题。这种方法能够保证规划所得作战方案是可执行的,飞行航线满足飞行器战技性能的约束,特别是多机协同对地攻击时,所有飞机都能按照规划的方案执行,而且作战效果最优或令人满意。而目前常见的方法是任务分配时使用直线距离或近似值来替代飞行航线,这种近似既无法保证规划方案的最优,还可能导致最终规划得到的作战方案存在冲突或超出飞机的能力,使得方案不可行。在航迹规划方面,本文总结了飞行器航迹规划问题的基本要素、约束条件、航迹规划的典型方法;重点分析了基于动态规划的航迹规划方法及其关键技术,提出了邻域网格点的的斜排列方式和消息的发送机制;提出了对航线进行水平修正的方法。仿真实验表明,本文的基于动态规划的航迹规划方法时间复杂度具有只与地理范围的大小有关的特点,对地形和地面威胁不敏感,而且该方法求解多任务点的航迹规划问题时比常见方法速度更快。在任务分配方面,本文对飞行器任务分配问题进行了建模,总结了常用的任务分配方法;重点介绍了基于多子群蚁群算法的任务分配方法及其关键技术,考虑被攻击目标的价值随时间动态变化的情况,提出了基于综合能力的子群选择策略、基于约束条件的动态候选集合策略、基于价值的状态转移策略、信息素的局部和全局更新机制,给出了算法的操作步骤。实验表明该任务分配方法有利于形成飞行器间协同合作的作战方案,而且在本文的约束条件下,当攻击目标数量较少时本文方法可以在短时间内规划出最优攻击方案,当攻击目标数量较多时则能在短时间内规划出次优的攻击方案。更多还原

【Abstract】 The capability of air-to-ground attacking is the decisive factor for Air Force to transform from positive defensive type to both offensive and defensive one. It has been attached great important attention by different countries all over the world. Therefore, studying the related technologies of air-to-ground attacking has a great significance to improve our air fighting force. This thesis mainly consider how to effectively use the aerial vehicles to implement air-to-ground attacking task when making the battle plan. The programming of air-to-ground attacking task includes two parts: flight route planning and task assignment. These two parts are not only mutual coupling interconnection but also very complex issues, this thesis particularly consider the practical situation that exists a variety of threats in the battle space, whice makes the difficulties of flight route planning and task assignment significantly increase.A approach for problem of integrated route planning and task assignment was proposed in this thesis,that is:applying route planning approach based on Dynamic Programming(DP) to search routes between any two task-points first,then applying task assignment approach based on the multi-ant-colony algorithm(MACA) to assign task for vehicles.So can deal with the constraints of the vehicles’capabilities and environment in the process of route planning,and deal with the constraints of tasks and cooperation in the process of task assignment.This approach can ensure the programming battle plan is executive,the flight routes conform to the vehicles’capabilities, especially that in the cooperative air-to-ground attacking task,every vehicle abides by the battle plan and the combat effectiveness is maximezed.The common approach at present is using straight line distance or it’s approximation to assign tasks, which can’t assure the battle plan is optimal,and maybe lead to an unexecutive plan due to incompatible battle plan or outreach of the vehicles’capabilities.In the terms of route planning,this thesis summarizes the basic elements、constraint conditions and typical approaches of vehicle’s route planning problem;mainly analyzed the route planning approach and it’s key technologies which based on DP, the slanting array of the frontier points and the mechanism of sending message were proposed; the horizontal improve method of routes was proposed. Simulation results show that the route planning approach based on dynamic programming presented in this thesis has a particular charater: it’s time complexity only related to the size of the geographic range,not sensitive to the terrain and the ground threats,and this approach is faster and the usual approach when sovling the route planning between multiple points.In the terms of task assignment, this thesis modeling the problem of vehicle’s task assignment,and summarize it’s usual sovling approach;mainly analyzed the task assignment approach and it’s key technologies based on MACA, consider the situation of target’s value reduce following the time gone,these were proposed: the selection strategy based on the integrative apability,the dynamic select-set strategy based on the constaints,the transform strategy based on values,the local and the global strategy of pheromone updating,and the operation steps of the algorithm was given.Simulation results show that the approach proposed in this thesis is beneficial to program a cooperative battle plan, and under the constraints listed, this approach can program the optimal battle plan in a short time when exits a little targets in the map,and the sub-optimal battle plan in a short time when exits a number of tartges in the map.更多还原