Simultaneous task assignment and path planning using mixed-integer linear programming and potential field method
Simultaneous task assignment and path planning using mixed-integer linear programming and potential field method
复制标题
DOI：
10.1109/iccas.2013.6704241
复制
发表时间：
2013-10
期刊：
影响因子：
--
通讯作者：
Seil An;H. Kim
中科院分区：
文献类型：
--
作者：
Seil An;H. Kim
研究了存在威胁情况下的任务分配和路径规划问题。一般的任务分配算法都是在任务分配阶段先获得直线路径，然后根据威胁进行路径规划，这会降低任务分配的性能和任务成功率。因此，本文提出了基于混合线性规划（MILP）和势场法的实时路径规划任务分配问题。实时路径规划通常需要大量的计算时间，但MILP简化了任务分配问题，因此实时路径规划可以应用于任务分配。提出了一种任务选择算法来补偿修正后的次优MILP。将MILP任务选择算法与在无威胁情况下获得最优解的蛮力搜索算法进行了比较。仿真结果表明，MILP任务选择算法在规避威胁和成功执行任务方面的优势。
This paper investigates a task assignment and path planning problem in a situation where threats exist. Common task assignment algorithms obtain straight line paths during the task assignment phase and then plan the paths considering the threats, which can decrease performance of task assignment and task success rate. Therefore, task assignment with real-time path planning is presented in this paper using Mixed-Integer Linear Programming(MILP) and potential field method. Real-time path planning usually requires amount of computational time but MILP simplifies the task assignment problem so real-time path planning can be applied to task assignment. Task selection algorithm is also presented to compensate the modified sub-optimal MILP. MILP task selection algorithm is compared with a brute-force search algorithm which obtains optimal solutions in situations without threats. Simulation results are presented to show the advantages of MILP task selection algorithm in avoiding threats and performing tasks successfully.