I’m writing my master thesis and i’ve developed a model for job shop schedule for an industrial application.
The problem exceeds the GAMS free license limitations, therefore i asked for a gurobi license and i’m now using gurobipy with python in order to be able to solve it.
The model that i’ve made with gurobipy works (kind of), but returns Solution count 0: Model is infeasible or unbounded and given my inexperience with both python and gurobi i think i must have made some kind of mistake while writing it.
Can someone with a full GAMS license double check if my GAMS model has a feasible solution? In that case i’ll know that there is something wrong with the gurobipy model and i’ll have a reference to fix it. I’ll PM you the problem in case.
There are reasons that licenses exist, but we do try and make it easy for certain groups to obtain access to “community licenses”. A community license is (from our website)…
“Users can request a free community license from sales@gams.com. The community license lets you generate and solve linear models (LP, MIP, and RMIP) that do not exceed 5000 variables and 5000 constraints. For all other model types the model cannot be larger than 2500 variables and 2500 constraints. Restrictions you experience with the demo license, i.e. additional limits enforced by some solvers, time-limitation of 12 months, and a recent GAMS system also apply to the community license. The free demo and community license cannot be combined with a professional license. These free demo and community licenses are for demonstration and evaluation but not for commercial and production work.”
This gives you a bit more room to build your model. I think that you will only find frustration if you continue on the road you are currently on.
Achristensen and Bussieck, thank you both for your responses.
I had no idea that GAMS had a community license with less restrictive problem dimension limitations.
Since the problem i’m using for testing uses a minuscule fraction of the real dataset i should have no problems running it with the community license, here’s the dimensions:
1101 rows, 274 columns and 3363 nonzeros
Variable types: 84 continuous, 190 integer (190 binary)
Non linear
When the model will be actually used in the real commercial application it will be used with gurobipy with its commercial license, but for now this possibility is so needed to help me debug the python code much faster.
Thanks!