arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.00260math.OCcs.ROcs.SYeess.SY
收藏

用符号最优控制解非线性推销员动力学的流动推销员问题

On the Solution of the Travelling Salesman Problem for Nonlinear Salesman Dynamics using Symbolic Optimal Control

Munich University of Applied Sciences(慕尼黑应用科学大学)

Alexander Weber, Alexander Knoll

摘要:

针对著名的流动推销员问题(TSP),当推销员的路径在连续状态空间和离散时间内演化,但具有任意(非线性)动态特性时,提出了一种求解TSP的算法.所提出的方法是基于符号控制的框架。通过这种方式,我们的方法返回一个可证明是正确的状态反馈控制器的基本覆盖规范,这是TSP离开了对路线的最优性的要求。此外,我们利用Lin-Kernighan-Helsgaun TSP求解器来优化整个路线的成本。两个例子,城市包裹递送任务和无人机侦察任务,极大地说明了所提出的启发式的强大功能。

英文摘要:

This paper proposes an algorithmic method to heuristically solve the famous Travelling Salesman Problem (TSP) when the salesman's path evolves in continuous state space and discrete time but with otherwise arbitrary (nonlinear) dynamics. The presented method is based on the framework of Symbolic Control. In this way, our method returns a provably correct state-feedback controller for the underlying coverage specification, which is the TSP leaving out the requirement for optimality on the route. In addition, we utilize the Lin-Kernighan-Helsgaun TSP solver to heuristically optimize the cost for the overall taken route. Two examples, an urban parcel delivery task and a UAV reconnaissance mission, greatly illustrate the powerfulness of the proposed heuristic.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑