Synthetic dataset used in "The maximum weighted submatrix coverage problem: A CP approach"

Synthetic dataset used in "The maximum weighted submatrix coverage problem: A CP approach".

Includes both the generated datasets as a zip archive and the python script used to generate them.

Each instance is composed of two files in the form

• XxY_K_O_0xN_AxB_Smatrix.tsv being the matrix to use. Each row on a separate line, with tab-separated cells.
• XxY_K_O_0xN_AxB_Ssolution.txt giving the implanted solution. One submatrix per line. Then two JSON arrays follow, separated by a tabulation. The first is the list of rows selected in the submatrix, the second the columns.

With:

• X and Y the size of the matrix
• K the number of submatrices in the implanted solution
• O the (minimum) overlap percentage of each submatrix
• N the sigma used for the background noise
• A and B the size of the implanted submatrices (subject to noise)