[Ipopt] Multiple Entries for Triplet matrix format

[Christian Wolf]

Hello Jon,

as I understand that correctly, if IPOPT detects a position in the 
matrix that you mention twice in the row and column vectors of the 
triplets it sums up the entries in the value vectors later on.

This might come in hand if you have a well structured problem (eg block 
structure) but some entries overlap with other blocks (think of coupling 
constraints).

This is how I understand that part of the documentation. If I am 
completely wrong please correct me, I am just trying to take advantage 
of this special feature in my next optimization problem formulation.

Chrees
Christian

Am 29.05.2015 um 12:57 schrieb Jonathan Gan:
>
> It is mentioned in the documentation that multiple entries might occur 
> for the triplet format for matrix representation for some nonzero 
> positions.  What is intended with "the values provided for those 
> positions are added"?
>
> Does it mean the value for that particular nonzero position with 
> multiple entries should be arithmetically added? Or should it be 
> overwritten with a later occurring entry?
>
> Cheers,
> Jon
>
>
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/ipopt/attachments/20150529/21c84a01/attachment.html>