On the integrality Gap of Small Asymmetric TSPs: A Polyhedral and Computational Approach

This folder contains the instances proposed in Table 2 and Table 3 of the paper "On the integrality Gap of Small Asymmetric TSPs: A Polyhedral and Computational Approach".

Table 3

In the table3 folder, the instances are provided in npy format.

The cost vector as well as the vertex attaining the optimal ASEP solution are provided

In practice, you can simply use the following code to read the instances:

X = np.load('table3/X_11.npy')
C = np.load('table3/C_11.npy')

Where X is the vertex and C is the cost vector.

The dimension of X is $n \times (n-1)$ where $n$ is the number of nodes in the graph. The values of the edges are here displayed in lexicographical order. The dimension of C is $n \times n$.

Table 4

The instances of Table 4 can be found in the table4 folder.

As explained in the paper, the ATSP instances have been transformed and rescaled according to be parsed by Concorde.

The original instances can be found in the table4/ATSP folder in the .atsp format, while the corresponding transformed instances are in the table4/STSP folder in the .tsp format.

Note that Concorde only deal with .tsp format.