I could not change Couenne's options in JuMP / CoinOptServices.js in a non linear model.
All my sources are updated: (Julia 0.6.2, JumP 0.18.1, CoinOptServices 0.3, AmplNLWriter 0.4)
m = Model(solver=AmplNLSolver(CoinOptServices.couenne,["couenne.allowable_gap=1"]))
The message is Unknown keyword "couenne.allowable_gap"
The options like allowable_gap are documented in couenne.opt
As the doc suggest, I've also tried to create a file (couenne.opt), that was placed tentatively in 2 differents folders: the same folder than source julia code with solver calls and the same folder than couenne.exe (bin) is, with no results.
File content: allowable_gap 1
For BonMin heuristics, the same syntax works and runs the model OK:
m = Model(solver=AmplNLSolver(CoinOptServices.bonmin,["bonmin.nlp_log_level=1"))
I could not change
Couenne
's options in JuMP / CoinOptServices.js in a non linear model. All my sources are updated: (Julia 0.6.2
,JumP 0.18.1
,CoinOptServices 0.3
,AmplNLWriter 0.4
)I've followed the instructons in AmplNLWriter.jl
m = Model(solver=AmplNLSolver(CoinOptServices.couenne,["couenne.allowable_gap=1"]))
The message is
Unknown keyword "couenne.allowable_gap"
The options like
allowable_gap
are documented in couenne.optAs the doc suggest, I've also tried to create a file (
couenne.opt
), that was placed tentatively in 2 differents folders: the same folder than source julia code with solver calls and the same folder thancouenne.exe
(bin) is, with no results.File content:
allowable_gap 1
For
BonMin
heuristics, the same syntax works and runs the model OK:m = Model(solver=AmplNLSolver(CoinOptServices.bonmin,["bonmin.nlp_log_level=1"))
However, the file
bonmin.opt
also fails