arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.00285cs.AIcs.LGcs.RO
收藏

iMTSP:用命令式学习求解最小-最大多旅行商问题

iMTSP: Solving Min-Max Multiple Traveling Salesman Problem with Imperative Learning

Purdue University(普渡大学);Carnegie Mellon University(卡内基梅隆大学);University at Buffalo(纽约州立大学布法罗分校)

Yifan Guo, Zhongqiang Ren, Chen Wang

AI总结:

本文提出基于命令式学习的双层优化框架iMTSP,通过分配网络分解MTSP并引入控制变量算法降低梯度方差,实现比强化学习基线快20%的收敛速度,在大规模问题上路径长度较Google OR-Tools最多缩短80%。

摘要:

本文考虑了最小-最大多旅行商问题(MTSP),其目标是为每个智能体找到一条路径,以共同访问所有城市,同时使最长路径的长度最小化。尽管 MTSP 已被广泛研究,但由于其 NP 难性质,为大规模问题获取近优解仍然具有挑战性。近期基于数据驱动的方法面临着需要难以获取的监督以及梯度估计高方差问题的挑战,导致收敛缓慢和解的高度次优。我们通过使用命令式学习(IL)的概念,将 MTSP 重新表述为双层优化问题来解决这些问题。这涉及引入一个分配网络,将 MTSP 分解为多个单智能体旅行商问题(TSP)。然后将这些 TSP 解中的最长路径用于自监督分配网络,从而产生了一个新的自监督、双层、端到端学习框架,我们称之为 imperative MTSP (iMTSP)。此外,为了解决优化过程中的高方差梯度问题,我们引入了一种基于控制变量的梯度估计算法。我们的实验表明,这些创新设计使我们的梯度估计器比先进的强化学习基线收敛快 20%,并且与 Google OR-Tools MTSP 求解器相比,找到的路径长度最多缩短 80%,尤其是在大规模问题中(例如 1000 个城市和 15 个智能体)。

英文摘要:

This paper considers a Min-Max Multiple Traveling Salesman Problem (MTSP), where the goal is to find a set of tours, one for each agent, to collectively visit all the cities while minimizing the length of the longest tour. Though MTSP has been widely studied, obtaining near-optimal solutions for large-scale problems is still challenging due to its NP-hardness. Recent efforts in data-driven methods face challenges of the need for hard-to-obtain supervision and issues with high variance in gradient estimations, leading to slow convergence and highly suboptimal solutions. We address these issues by reformulating MTSP as a bilevel optimization problem, using the concept of imperative learning (IL). This involves introducing an allocation network that decomposes the MTSP into multiple single-agent traveling salesman problems (TSPs). The longest tour from these TSP solutions is then used to self-supervise the allocation network, resulting in a new self-supervised, bilevel, end-to-end learning framework, which we refer to as imperative MTSP (iMTSP). Additionally, to tackle the high-variance gradient issues during the optimization, we introduce a control variate-based gradient estimation algorithm. Our experiments showed that these innovative designs enable our gradient estimator to converge 20% faster than the advanced reinforcement learning baseline and find up to 80% shorter tour length compared with Google OR-Tools MTSP solver, especially in large-scale problems (e.g. 1000 cities and 15 agents).

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。

↑