0

I have been searching quite a while without success, getting headaches ;-).

Consider the TSP (traveling salesman problem), with a list of nodes 0, 1....n-1 BUT:

  • trip must start at 0 and end at 0

  • there is just a known distance between all nodes

  • trip must be a "bitonic": id est visit increasing numbered nodes, n-1, decreasing numbered nodes (remaining ones of course).

I am trying hard to get the recursive formula :

o(n,p) = f { all o(k,l) with k <= n l <=p } etc...

I want a DP (dynamic programming) approach. Besides, I have already successfully implemented (in python) the (kind of) brute force approach.

Any suggestion ? Thanks!!!

Jeremie
  • 399
  • 4
  • 11
  • You need to read https://medium.com/basecs/speeding-up-the-traveling-salesman-using-dynamic-programming-b76d7552e8dd – timday Jul 27 '19 at 18:41
  • thanks timday, I have already read the (very good) page you mention. I have already implemented and understood a TSP with DP. But I am still stuck on the variant exposed above. – Jeremie Jul 27 '19 at 18:58
  • 2
    Possible duplicate of [How to compute optimal paths for traveling salesman bitonic tour?](https://stackoverflow.com/questions/874982/how-to-compute-optimal-paths-for-traveling-salesman-bitonic-tour) – Tassle Jul 27 '19 at 20:35
  • Yep, duplicate. With the explanations of the link provided by Tassle I was able to figure it out. thanks! – Jeremie Jul 27 '19 at 21:33

0 Answers0