Trouble running the hydro example

[bartimore]

Dear All,

First I want to thank you for the development of this package. It seems to be exactly what I need.
The problem I want to solve is quite similar to the hydro problem, so I am trying to understand it first.

Though I managed to run hydro_ef.py, I have trouble running the cylinder scripts from the command line.

Passing the branching-factor
When trying to run e.g.:
python hydro_cylinders.py --branching-factors 3 3 --max-iterations 20 --solver-name "gurobi_direct" --default-rho 1
I get the error message:
farmer_cylinders: error: unrecognized arguments: 3
This error seems to apply to the second "3". The -h tells me to provide branching-factors as: "Spaces delimited branching factors (e.g., 2 2)", but this does not seem to work for me. Do you know what I am doing wrong?

No incumbent solution
More worryingly, if I run:
python hydro_cylinders_pysp.py --max-iterations 50 --default-rho 1.0 --solver-name "gurobi_direct"
(which "ignores the branching factors" (how?)) I get no incumbent solution (see below). Is it true that this should be the case? I didn't make any changes to the model. Starting from an example that does not solve the model, makes it very hard to start building upon it.

Thank you in advance!
`
PS C:\Users\s147670\PycharmProjects\MultiStage\mpi-sppy\examples\hydro> python hydro_cylinders_pysp.py --max-iterations 50 --default-rho 1.0 --solver-name "gurobi_direct"
[ 0.00] Initializing mpi-sppy
[ 1.79] Initializing SPBase
[ 1.85] Initializing PHBase
[ 1.86] Starting spcomm.main()
[ 1.86] Creating solvers
[ 1.88] Entering solve loop in PHBase.Iter0
[ 2.00] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap
[ 2.00] 1 * 175.0168 inf inf% inf
[ 2.06] 2 175.0168 inf inf% inf
[ 2.12] 3 175.0168 inf inf% inf
[ 2.18] 4 175.0168 inf inf% inf
[ 2.24] 5 175.0168 inf inf% inf
[ 2.30] 6 175.0168 inf inf% inf
[ 2.36] 7 175.0168 inf inf% inf
[ 2.42] 8 175.0168 inf inf% inf
[ 2.48] 9 175.0168 inf inf% inf
[ 2.54] 10 175.0168 inf inf% inf
[ 2.59] 11 175.0168 inf inf% inf
[ 2.65] 12 175.0168 inf inf% inf
[ 2.71] 13 175.0168 inf inf% inf
[ 2.77] 14 175.0168 inf inf% inf
[ 2.83] 15 175.0168 inf inf% inf
[ 2.89] 16 175.0168 inf inf% inf
[ 2.95] 17 175.0168 inf inf% inf
[ 3.01] 18 175.0168 inf inf% inf
[ 3.07] 19 175.0168 inf inf% inf
[ 3.14] 20 175.0168 inf inf% inf
[ 3.19] 21 175.0168 inf inf% inf
[ 3.25] 22 175.0168 inf inf% inf
[ 3.31] 23 175.0168 inf inf% inf
[ 3.37] 24 175.0168 inf inf% inf
[ 3.42] 25 175.0168 inf inf% inf
[ 3.48] 26 175.0168 inf inf% inf
[ 3.54] 27 175.0168 inf inf% inf
[ 3.59] 28 175.0168 inf inf% inf
[ 3.65] 29 175.0168 inf inf% inf
[ 3.71] 30 175.0168 inf inf% inf
[ 3.77] 31 175.0168 inf inf% inf
[ 3.83] 32 175.0168 inf inf% inf
[ 3.88] 33 175.0168 inf inf% inf
[ 3.94] 34 175.0168 inf inf% inf
[ 4.00] 35 175.0168 inf inf% inf
[ 4.06] 36 175.0168 inf inf% inf
[ 4.11] 37 175.0168 inf inf% inf
[ 4.17] 38 175.0168 inf inf% inf
[ 4.23] 39 175.0168 inf inf% inf
[ 4.29] 40 175.0168 inf inf% inf
[ 4.34] 41 175.0168 inf inf% inf
[ 4.40] 42 175.0168 inf inf% inf
[ 4.46] 43 175.0168 inf inf% inf
[ 4.57] 45 175.0168 inf inf% inf
[ 4.63] 46 175.0168 inf inf% inf
[ 4.69] 47 175.0168 inf inf% inf
[ 4.74] 48 175.0168 inf inf% inf
[ 4.80] 49 175.0168 inf inf% inf
[ 4.86] 50 175.0168 inf inf% inf
[ 4.86] Reached user-specified limit=50 on number of PH iterations
[ 4.86] Hub algorithm PH complete, waiting for spoke finalization
[ 4.86] Statistics at termination
[ 4.86] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap
[ 4.86] 50 175.0168 inf inf% inf
[ 4.86] Windows freed
BestInnerBound=inf and BestOuterBound=175.0168013257502
[ 4.86] No incumbent solution available to write!
[ 4.86] No incumbent solution available to write!

`

[bartimore]

I just noticed that I encounter the same problems if I run the farmer_cylinder.py script. The examples from: https://mpi-sppy.readthedocs.io/en/latest/examples.html do work as desired.

PS C:\Users\s147670\PycharmProjects\MultiStage\mpi-sppy\examples\farmer> python farmer_cylinders.py --num-scens 3 --crops-mult 1 --default-rho 1 --solver-name "gurobi_direct" --max-iterations 10

It seems like I am making the same mistake twice, but I don't know what means I can use to figure out what is going wrong. I think it would be very helpful to show in the examples how to run the scripts from the command line to get working solutions.

[ 0.00] Initializing mpi-sppy
[ 1.53] Initializing SPBase
[ 1.53] Initializing PHBase
[ 1.53] Starting spcomm.main()
[ 1.53] Creating solvers
[ 1.55] Entering solve loop in PHBase.Iter0
[ 1.59] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap
[ 1.59] 1 * -115405.5556 inf inf% inf
[ 1.61] 2 -115405.5556 inf inf% inf
[ 1.62] 3 -115405.5556 inf inf% inf
[ 1.64] 4 -115405.5556 inf inf% inf
[ 1.66] 5 -115405.5556 inf inf% inf
[ 1.68] 6 -115405.5556 inf inf% inf
[ 1.70] 7 -115405.5556 inf inf% inf
[ 1.72] 8 -115405.5556 inf inf% inf
[ 1.74] 9 -115405.5556 inf inf% inf
[ 1.76] 10 -115405.5556 inf inf% inf
[ 1.76] Reached user-specified limit=10 on number of PH iterations
Arbitrary sanity checks:
SUGAR_BEETS0 for scenario scen0 is 236.19106321346567
FirstStageCost for scenario scen0 is 108538.4744787988
[ 1.76] Hub algorithm PH complete, waiting for spoke finalization
[ 1.76] Statistics at termination
[ 1.76] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap
[ 1.76] 10 -115405.5556 inf inf% inf
[ 1.76] Windows freed
[ 1.76] No incumbent solution available to write!
[ 1.76] No incumbent solution available to write!
[ 1.76] No incumbent solution available to write!

[DLWoodruff]

The command you are running is working as expected, it only has a hub cylinder (no bounds cylinders). Assuming you have MPI and mpi4py installed properly (see the readme and when in doubt, go the anaconda route): mpiexec -np 3 python -m mpi4py farmer_cylinders.py --num-scens 3 --crops-mult 1 --default-rho 1 --solver-name "gurobi_direct" --max-iterations 10 --lagrangian --xhatshuffle (see simple.bash in the farmer example for more command lines)

…

On Fri, Aug 16, 2024 at 5:25 AM Bart van der Holst ***@***.***> wrote: I just noticed that I encounter the same problems if I run the farmer_cylinder.py script. The examples from: https://mpi-sppy.readthedocs.io/en/latest/examples.html <http://url> do work as desired. PS C:\Users\s147670\PycharmProjects\MultiStage\mpi-sppy\examples\farmer> python farmer_cylinders.py --num-scens 3 --crops-mult 1 --default-rho 1 --solver-name "gurobi_direct" --max-iterations 10 [ 0.00] Initializing mpi-sppy [ 1.53] Initializing SPBase [ 1.53] Initializing PHBase [ 1.53] Starting spcomm.main() [ 1.53] Creating solvers [ 1.55] Entering solve loop in PHBase.Iter0 [ 1.59] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap [ 1.59] 1 * -115405.5556 inf inf% inf [ 1.61] 2 -115405.5556 inf inf% inf [ 1.62] 3 -115405.5556 inf inf% inf [ 1.64] 4 -115405.5556 inf inf% inf [ 1.66] 5 -115405.5556 inf inf% inf [ 1.68] 6 -115405.5556 inf inf% inf [ 1.70] 7 -115405.5556 inf inf% inf [ 1.72] 8 -115405.5556 inf inf% inf [ 1.74] 9 -115405.5556 inf inf% inf [ 1.76] 10 -115405.5556 inf inf% inf [ 1.76] Reached user-specified limit=10 on number of PH iterations Arbitrary sanity checks: SUGAR_BEETS0 for scenario scen0 is 236.19106321346567 FirstStageCost for scenario scen0 is 108538.4744787988 [ 1.76] Hub algorithm PH complete, waiting for spoke finalization [ 1.76] Statistics at termination [ 1.76] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap [ 1.76] 10 -115405.5556 inf inf% inf [ 1.76] Windows freed [ 1.76] No incumbent solution available to write! [ 1.76] No incumbent solution available to write! [ 1.76] No incumbent solution available to write! — Reply to this email directly, view it on GitHub <#409 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AD4VTBEXGXZ4FZMPPOZ7AWTZRXVURAVCNFSM6AAAAABMT5YYZSVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTAMZVHA2DEMY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[bartimore]

Thanks a lot! This worked for me!

[DLWoodruff]

One seems to need quotes around the 3 3 as in "3 3" Also, if you want bounds (or a solution), you need to run as in the file demo.bash (except that it doesn't have the necessary quotes around 3 3). Dave

…

On Fri, Aug 16, 2024 at 4:59 AM Bart van der Holst ***@***.***> wrote: Dear All, First I want to thank you for the development of this package. It seems to be exactly what I needed. The problem I want to solve is quite similar to the hydro problem, so I am trying to understand it first. Though I managed to run hydro_ef.py, I do some trouble with running the cylinder scripts from the command line. **Passing the branching-factors" When trying to run e.g.: python hydro_cylinders.py --branching-factors 3 3 --max-iterations 20 --solver-name "gurobi_direct" --default-rho 1 I get the error message: farmer_cylinders: error: unrecognized arguments: 3 This error seems to apply to the second "3". The -h tells me to provide branching-factors as: "Spaces delimited branching factors (e.g., 2 2)", but this does not seem to work for me. Do you know what I am doing wrong? *No incumbent solution* More worryingly, if I run: python hydro_cylinders_pysp.py --max-iterations 50 --default-rho 1.0 --solver-name "gurobi_direct" (which "ignores the branching factors" (how?)) I get no incumbent solution (see below). Is it true that this should be the case? I didn't make any changes to the model. Starting from an example that does not solve the model, makes it very hard to start building upon it. I thank you in advance! ` PS C:\Users\s147670\PycharmProjects\MultiStage\mpi-sppy\examples\hydro> python hydro_cylinders_pysp.py --max-iterations 50 --default-rho 1.0 --solver-name "gurobi_direct" [ 0.00] Initializing mpi-sppy [ 1.79] Initializing SPBase [ 1.85] Initializing PHBase [ 1.86] Starting spcomm.main() [ 1.86] Creating solvers [ 1.88] Entering solve loop in PHBase.Iter0 [ 2.00] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap [ 2.00] 1 * 175.0168 inf inf% inf [ 2.06] 2 175.0168 inf inf% inf [ 2.12] 3 175.0168 inf inf% inf [ 2.18] 4 175.0168 inf inf% inf [ 2.24] 5 175.0168 inf inf% inf [ 2.30] 6 175.0168 inf inf% inf [ 2.36] 7 175.0168 inf inf% inf [ 2.42] 8 175.0168 inf inf% inf [ 2.48] 9 175.0168 inf inf% inf [ 2.54] 10 175.0168 inf inf% inf [ 2.59] 11 175.0168 inf inf% inf [ 2.65] 12 175.0168 inf inf% inf [ 2.71] 13 175.0168 inf inf% inf [ 2.77] 14 175.0168 inf inf% inf [ 2.83] 15 175.0168 inf inf% inf [ 2.89] 16 175.0168 inf inf% inf [ 2.95] 17 175.0168 inf inf% inf [ 3.01] 18 175.0168 inf inf% inf [ 3.07] 19 175.0168 inf inf% inf [ 3.14] 20 175.0168 inf inf% inf [ 3.19] 21 175.0168 inf inf% inf [ 3.25] 22 175.0168 inf inf% inf [ 3.31] 23 175.0168 inf inf% inf [ 3.37] 24 175.0168 inf inf% inf [ 3.42] 25 175.0168 inf inf% inf [ 3.48] 26 175.0168 inf inf% inf [ 3.54] 27 175.0168 inf inf% inf [ 3.59] 28 175.0168 inf inf% inf [ 3.65] 29 175.0168 inf inf% inf [ 3.71] 30 175.0168 inf inf% inf [ 3.77] 31 175.0168 inf inf% inf [ 3.83] 32 175.0168 inf inf% inf [ 3.88] 33 175.0168 inf inf% inf [ 3.94] 34 175.0168 inf inf% inf [ 4.00] 35 175.0168 inf inf% inf [ 4.06] 36 175.0168 inf inf% inf [ 4.11] 37 175.0168 inf inf% inf [ 4.17] 38 175.0168 inf inf% inf [ 4.23] 39 175.0168 inf inf% inf [ 4.29] 40 175.0168 inf inf% inf [ 4.34] 41 175.0168 inf inf% inf [ 4.40] 42 175.0168 inf inf% inf [ 4.46] 43 175.0168 inf inf% inf [ 4.57] 45 175.0168 inf inf% inf [ 4.63] 46 175.0168 inf inf% inf [ 4.69] 47 175.0168 inf inf% inf [ 4.74] 48 175.0168 inf inf% inf [ 4.80] 49 175.0168 inf inf% inf [ 4.86] 50 175.0168 inf inf% inf [ 4.86] Reached user-specified limit=50 on number of PH iterations [ 4.86] Hub algorithm PH complete, waiting for spoke finalization [ 4.86] Statistics at termination [ 4.86] Iter. Best Bound Best Incumbent Rel. Gap Abs. Gap [ 4.86] 50 175.0168 inf inf% inf [ 4.86] Windows freed BestInnerBound=inf and BestOuterBound=175.0168013257502 [ 4.86] No incumbent solution available to write! [ 4.86] No incumbent solution available to write! ` — Reply to this email directly, view it on GitHub <#409>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AD4VTBA46ALEIQETJK2DBJDZRXSTXAVCNFSM6AAAAABMT5YYZSVHI2DSMVQWIX3LMV43ERDJONRXK43TNFXW4OZXGA2TSNZRGY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[bartimore]

Thanks David! This worked for me!