[Osi-tickets] [COIN-OR Open Solver Interface] #28: getDualRays maxNumRays, but only 1 in CLP/CPX

COIN-OR Open Solver Interface coin-trac at coin-or.org
Fri Jul 18 01:20:29 EDT 2008


#28: getDualRays maxNumRays, but only 1 in CLP/CPX
-----------------------------+----------------------------------------------
  Reporter:  mgalati         |       Owner:  mjs     
      Type:  enhancement     |      Status:  new     
  Priority:  minor           |   Milestone:          
 Component:  Osi Base Class  |     Version:  0.95.2  
Resolution:                  |    Keywords:  dual ray
-----------------------------+----------------------------------------------
Changes (by lou):

  * priority:  major => minor

Old description:

> In both OsiClp and OsiCpx, getDualRays takes arg=maxNumRays, but only
> returns 1 dual ray. Can the solvers provide more than one? If so, can
> this be implemented (this would be nice)? If not, should there be a
> maxNumRays argument at all? It seems misleading.

New description:

 In both !OsiClp and !OsiCpx, getDualRays takes arg=maxNumRays, but only
 returns 1 dual ray. Can the solvers provide more than one? If so, can this
 be implemented (this would be nice)? If not, should there be a maxNumRays
 argument at all? It seems misleading.

Comment:

 Just finished crafting the dylp equivalent, so I can offer some
 observations. It's not particularly hard to find all rays at the vertex
 where unboundedness is detected. Finding other rays, if they exist,
 amounts to walking the vertices of the polytope. The computational cost
 would be fairly high.

-- 
Ticket URL: <https://projects.coin-or.org/Osi/ticket/28#comment:1>
COIN-OR Open Solver Interface <http://projects.coin-or.org/Osi>
An API for a variety of LP and MIP solvers (and more).



More information about the Osi-tickets mailing list