Talk:Vehicle routing problem

How come the Clarke & Wright savings heuristic not being mentioned anywhere on the English Wikipedia? Was quite hard finding this page. --Blackstab-- (talk) 15:31, 26 November 2010 (UTC)

Isn't this problem the same as the Job-shop problem, with machines being vehicles and jobs being customers?--130.115.76.138 19:40, 3 February 2007 (UTC)

Not true. I've added a section for this (with references).Jiuguang Wang (talk) 17:20, 18 April 2008 (UTC)

The ``Why is it hard? section seems silly to me. That is, it occupies about 40% of the article and it really is just a basic tutorial on combinatorial explosion. Surely there are more interesting facets of vehicle routing that can be discussed instead of this filler. — Preceding unsigned comment added by 129.97.166.49 (talk) 19:31, 20 June 2013 (UTC)

I agree. Deleted. AustinBuchanan (talk) 00:44, 19 September 2013 (UTC)

Why is it hard part, is the most useful part of this wiki, since regular people like me with a standard academic background including operations research try to approach the problem the "simplex" way, since the problem itself doesnt look like NP-Hard. Deleting is not the way to go, adding more content could be in my opinion. — Preceding unsigned comment added by 195.46.135.130 (talk) 11:13, 19 September 2013 (UTC)

Vehicle flow formulations

I think the notation needs to be explained in the vehicle flow formulation section. Cij is the cost of going from node i to node j, Xij is a boolean that is 1 if the edge ij is part of the solution and 0 otherwise. Node 0 is the depot node, K is the number of vehicles. All this needs to be explained. 118.211.63.81 (talk) 01:49, 12 August 2018 (UTC)

good point feel free to add. Lady-shirakawa (talk) 09:37, 21 August 2018 (UTC)

Did anyone check the source for the MTZ constraints? Because (a) the explanation is a bit confusing and (b) the two constraints reflect opposite intuitions: in the first u grows (reflecting additional free capacity during a delivery), but in the second, it is restricted by the demand of a customer, which somehow contradicts the constraint before. Also, there are no restrictions coming from/going to the depot node. So either it is more than just a bit unclear, or it is plainly wrong.

84.138.65.139 (talk) 07:37, 1 September 2020 (UTC)

I didn't have the referenced paper but I 'm convinced too there is a typo. The additionnal constraints should read :

(**) u_i - u_j >= d_j - C(1 - x_{ij})

and not :

(*) u_j - u_i >= d_j - C(1 - x_{ij})

Bruno P. 193.50.40.239 (talk) 15:43, 1 February 2023 (UTC)