Qkb
Qkb
Hi, the code of backbone Att-GCRN is baed on Joshi-etal, 2019, and also I have add training tricks and done some modification about the implementation of attention. In our paper,...
Hi, i have not reproduced this problem when runnning the related codes. Please check the hyper-parameter of KNN.
Hi, taking more care of the format of your instances.
Hi, this result of table is average length of predicted solution in testset.
Hi, have you recompiled your project after modify related codes?
Hi, how long did your MCTS run? The MCTS also is learning methods. If the running time of MCTS is long enough, nearly optimized solution will be searched.
Hi, all experiments were tested in GPU [in this paper.](https://arxiv.org/abs/2012.10658) The running time of MCTS is longer [in another paper](https://openreview.net/forum?id=ByxtHCVKwB).
Hi @bywbilly , I have not encountered this problem during the process of experiment.
Hi， 1. As described in the paper, simply adjusting the parameters in the clustering phase can solve the Traveling Salesman Problem (TSP) with any number of nodes. 2. Yes, you...