While it may look silly to ask Clp to solve such a trivial problem, this actually happens as part of a larger algorithms (see here https://github.com/blegat/StochasticDualDynamicProgramming.jl) so it wasn't generated by hand and the fact that Clp fails on such simple problems makes it unusable for the whole algorithm :(
Moved from TRAC as part of COIN fORgery:
Dear Clp maintainers,
For the trivial problem (see attachment for an mps):
Minimize x0 + x1 Subject To
Bounds
x1 >= 0 x1 free
End
Clp correctly detects it to be unbounded but returns NULL when asked for the Unbounded ray (see here https://github.com/JuliaOpt/Clp.jl/pull/20).
While it may look silly to ask Clp to solve such a trivial problem, this actually happens as part of a larger algorithms (see here https://github.com/blegat/StochasticDualDynamicProgramming.jl) so it wasn't generated by hand and the fact that Clp fails on such simple problems makes it unusable for the whole algorithm :(
Attachment: https://projects.coin-or.org/Clp/attachment/ticket/82/ticket82.mps Trivial unbounded problem for which the unbounded ray is not found