Re: Discussion regarding Mr. Diabys algorithm


Radoslaw Hofman wrote:
Uzytkownik <moustapha.diaby@xxxxxxxxxxxxxxxxxx> napisal w wiadomosci
news:1163626104.127115.292010@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx


In summary I must conclude that you haven't written in this document
anything not written by you before. We all expect that you will stop writing
"[...] is impossible" but show us all WHY. Especially that I gave you
pseudo-math explanation of mine objection above (see 2). You didn't prove
that:


What I mean is this:

Consider arc (5, 4, 6) in your "valley" B of Figure 13. It is
*impossible* to have a set of positive y variables (
y(5,4,6),(i(s),s,j(s)), s = 4, ..., 23) along with a set of positive
z-variables (z(5,4,6),(i(s),s,j(s)),(i(t),t,j(t))), s = 5 to 22; t =s+1
to 23) such that they satisfy all the constraints of my model or
Proposition 2 in my paper.

Please, think about it a little!

How would flow on this arc "travel" from stage 4 to stage 23 without
"revisiting" any of the only 6 levels in your "valley" B, when there
are no "back-flows" allowed as stipulated in constraints 2.14, and as
you have already said there are none in this usenet?

I think it would be helpful if you can show one set of y and z-values
for arc (5,4,6) from your "counter-example" that satisfies all the
constraints of my model. (Forget everything else.)

//MD

.

Relevant Pages

  • Re: Discussion regarding Mr. Diabys algorithm
    ... The reason is that the variables are not ... "fly" from arc to arc. ... Flow must be connected (required by constraints 2.8); ...
    (comp.theory)
  • Re: P=NP: Linear Programming Formulation of the TSP
    ... > Any feasible solution to IPBAR can be generated as a sum ... a "level," and of yas the contribution of (the flow on) arc ... stipulate that an arc cannot contribute to any "downstream" arc ... stage 1 must span the set of stages because of constraints 2.15. ...
    (comp.theory)
  • Re: P=NP: Linear Programming Formulation of the TSP
    ... I hope you appreciate my efforts to> give you an intuitive sense of what the mathematics prove... ... Hence this amount does not "return" to either cities i or> j because of constraints 2.22. ... > Now assume the total flow on an arbitrary arc, y, is> positive. ...
    (comp.theory)
  • Re: Discussion regarding Mr. Diabys algorithm
    ... The reason is that the variables are not ... "fly" from arc to arc. ... Flow must be connected (required by constraints 2.8); ...
    (comp.theory)
  • Re: Discussion regarding Mr. Diabys algorithm
    ... are trees that are really huge and cannot be therefore, ... because of the relations between the z's and y's and constraints 2.14, ... In Mr. Hofman's "counter-example," either flow from a given arc ... conservation, Proposition 2 of my paper is true. ...
    (comp.theory)