[Coin-discuss] sbb assertion failure

[Jesper Hansen]

I did an MPS file of the input to sbb and did the same operations 
following, but now Sbb works fine. Strange?! Makes it a bit tricky to 
investigate...

Anyone interested in the code and mps file can mail me.

'Jesper

P.S. here is the clp/sbb output:

Clp0000I Optimal - objective value 1.6927e+06
Sbb0009I Objective coefficients multiple of 0.0031746
Sbb0013I At root node, 0 cuts changed objective from 1.6927e+06 to 
1.69295e+06 in 1 passes
Sbb0014I Cut generator 0 (Probing) - 2 row cuts (0 active), 6 column cuts 
- new frequency is -100
Sbb0014I Cut generator 1 (Gomory) - 0 row cuts (0 active), 0 column cuts - 
new frequency is -100
Sbb0014I Cut generator 2 (Knapsack) - 0 row cuts (0 active), 0 column cuts 
- new frequency is -100
Sbb0014I Cut generator 3 (OddHole) - 0 row cuts (0 active), 0 column cuts 
- new frequency is -100
Sbb0004I Integer solution of 1.69295e+06 found after 0 iterations and 0 
nodes
Sbb0001I Search completed - best objective 1.69295e+06, took 0 iterations 
and 0 nodes
Sbb0004I Integer solution of 1.69295e+06 found after 0 iterations and 0 
nodes  

The cut-off value is 1.69295e+06.

On Wed, 28 Jan 2004, Lou Hafer wrote:



> Jesper,
> 
> 	Yan is correct as to the why. We'll need an example to pursue the how
> of it. Can you send me an example that triggers the assertion? Probably best to 
> do it off list, unless it's trivially small.
> 	
> 							Lou
> 
> _______________________________________________
> Coin-discuss mailing list
> Coin-discuss at www-124.ibm.com
> http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-discuss
> 

-- 
Jesper Hansen, mailto:Jesper at carmenconsulting.com
Carmen Consulting, Købmagergade 53, Copenhagen K, Denmark 
Tel.: +45 33 17 91 92    Mobil: +45 61 33 29 69
http://www.carmenconsulting.com