Skip to content

Use L-BFGS if the Hessian of the Lagrangian is not available #147

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
amontoison wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Use L-BFGS if the Hessian of the Lagrangian is not available #147

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion Project.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@ SolverCore = "ff4d7338-4cf1-434d-91df-b86cb86fb843"
ADNLPModels = "0.8.13"
KNITRO = "1.1.0"
LinearAlgebra = "1.10"
NLPModels = "0.21.5"
NLPModels = "0.21.6"
NLPModelsModifiers = "0.7"
SolverCore = "0.3"
Test = "1.10"
Expand Down
13 changes: 13 additions & 0 deletions docs/src/tutorial.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -200,3 +200,16 @@ contour!(range(18, 60, step=0.1), range(40_000, 180_000, step=1.0), f, levels=[0
```

![](assets/ex3.png)

## Example: Using L-BFGS (limited-memory) Hessian approximation with KNITRO

You can call KNITRO with the L-BFGS Hessian approximation by passing the following options:

```julia
stats_knitro = knitro(nlp,
hessopt=6,
lmsize=10)
```

This will use the L-BFGS method for Hessian approximation with a history size of 10.
Note that these options are used by default in `NLPModelsKnitro.jl` if `nlp.meta.hess_available` is `false`.
18 changes: 15 additions & 3 deletions src/nlp.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ function KnitroSolver(
linear_api::Bool = true,
kwargs...,
)
@assert nlp.meta.grad_available && (nlp.meta.ncon == 0 || nlp.meta.jac_available)
n, m = nlp.meta.nvar, nlp.meta.ncon

kc = KNITRO.KN_new()
Expand Down Expand Up @@ -91,7 +92,12 @@ function KnitroSolver(
else
jrows, jcols = jac_structure(nlp)
end
hrows, hcols = hess_structure(nlp)

if nlp.meta.hess_available
hrows, hcols = hess_structure(nlp)
else
hrows, hcols = Int[], Int[]
end

# define evaluation callback
function evalAll(kc, cb, evalRequest, evalResult, userParams)
Expand Down Expand Up @@ -175,8 +181,14 @@ function KnitroSolver(
hessIndexVars2 = convert(Vector{Int32}, hrows .- 1),
)

# specify that we are able to provide the Hessian without including the objective
KNITRO.KN_set_int_param(kc, KNITRO.KN_PARAM_HESSIAN_NO_F, KNITRO.KN_HESSIAN_NO_F_ALLOW)
# Use L-BFGS if the sparse hessian of the Lagrangian is not available
if !nlp.meta.hess_available
KNITRO.KN_set_int_param_by_name(kc, "hessopt", 6)
KNITRO.KN_set_int_param_by_name(kc, "lmsize", 10)
else
# specify that we are able to provide the Hessian without including the objective
KNITRO.KN_set_int_param(kc, KNITRO.KN_PARAM_HESSIAN_NO_F, KNITRO.KN_HESSIAN_NO_F_ALLOW)
end

# pass options to KNITRO
for (k, v) in kwargs
Expand Down
Loading