[Cbc] refusing feasible solutions

[John Forrest]

Pietro,

checkSolution now virtual (in trunk)

Regards

John

On 03/09/12 05:05, Pietro Belotti wrote:
> John,
>
> apologies for the late reply. I am looking at the Bonmin code to see 
> where the solvercharacteristics makes a difference. As for the virtual 
> function, I guess I overlooked some library version and posted the 
> wrong problem: now, declaring a virtual CbcModel::checkSolution() 
> allows me to perform my own check in my model class, and Cbc uses only 
> the accepted solutions as valid. It would then be desirable to have 
> that function virtual in the repository.
>
> Thank you very much,
> Pietro
>
> -- 
> Pietro Belotti
> Dept. of Mathematical Sciences
> Clemson University
> email: pbelott at clemson.edu
> phone: 864-656-6765
> web:   http://myweb.clemson.edu/~pbelott
>
> On Thu, 26 Jul 2012, John Forrest wrote:
>
>> Pietro,
>>
>> When I run those two .lp files with standalone solver it works fine.
>>
>> The solvercharacteristics stuff was put in for Pierre Bonami and 
>> Bonmin - you could take a look at Bonmin code.
>>
>> However I think it will be cleaner to make a Cbc function virtual so -
>>
>> going back to your original code, where does it find the heuristic 
>> solution? can you see why it doesn't call a useful function?
>>
>> John
>> _______________________________________________
>> Cbc mailing list
>> Cbc at list.coin-or.org
>> http://list.coin-or.org/mailman/listinfo/cbc
>
>