Re: [Help-glpk] Preprocessing - Duplicate Row Removal

[Andrew Makhorin]

Hi Xypron,

> I have updated the implementation of duplicate row removal, to correctly set
> the status and
> the dual values.
> http://old.nabble.com/file/p30143969/glpnpp06.c glpnpp06.c 
> 
> The current implementation has a worst case time complexity of
> O(m^2 * n log(n)) where m the number of rows and n the number of columns)

Thank you very much for your contribution. I need a time to test your
code. (Just a request: please do not use tabs and lines longer than 72
chars in source files.)

I think to introduce a preprocessor option (or maybe a command-line
option in glpsol) to enable/disable new untested and experimental
features added to glpk. 

> 
> I am considering to implement the algorithm from
> R. E. Bixby and D. K. Wagner
> A Note on Detecting Simple Redundancies in Linear Systems,
> Operations Research Letters, vol 6, p 15-17, 1987
>  ( also Technical Report 86-14, June 1986,
> http://www.caam.rice.edu/caam/trs/86/TR86-14.pdf )
> 

It would be great!


Best regards,

Andrew Makhorin