skantz / tsp Goto Github PK

Bug: the mst code is really bad.

Expected behavior: the mst code is not bad.

Proposed fix: re-write everything to achieve a perfect score in pylint.

We need to write an MST report.

• Theory background
• Algorithm
• Local experiments

etc

Bug: we don't have a new tsp solution.
Expected behavior: we have a new tsp solution. In rust.

To write a good report, it would be good to have local tests for mst. We could compare our algorithm to the ground truth of Kruskal's mst. For which graph parameters or families does our algorithm give an accurate result?

We have the second best solution on https://kth.kattis.com/problems/kth.avalg.approxspanningforest/statistics

Can we achieve number 1?

To reproduce:
run mst/mst_local.ipynb and wait for approx. 5 min

Bug: we don̈́' have a tsp report.
Expected behavior: we have a longer tsp report.