bench: new test on the pendulum with MS and UnoSolver

Author: franckgaga

benchmark/3_bench_predictive_control.jl

@@ -318,7 +318,13 @@ optim = JuMP.Model(MadNLP.Optimizer, add_bridges=false)
 transcription = SingleShooting()
 nmpc_madnlp_ss = NonLinMPC(estim; Hp, Hc, Mwt, Nwt, Cwt, optim, transcription)
 nmpc_madnlp_ss = setconstraint!(nmpc_madnlp_ss; umin, umax)
-JuMP.unset_time_limit_sec(nmpc_madnlp_ss.optim) 
+JuMP.unset_time_limit_sec(nmpc_madnlp_ss.optim)
+
+optim = JuMP.Model(()->UnoSolver.Optimizer(preset="filtersqp"), add_bridges=false)
+transcription, hessian, oracle = MultipleShooting(), true, true
+nmpc_uno_ms_hess = NonLinMPC(estim; Hp, Hc, Mwt, Nwt, Cwt, optim, transcription, hessian, oracle)
+nmpc_uno_ms_hess = setconstraint!(nmpc_uno_ms_hess; umin, umax)
+JuMP.unset_time_limit_sec(nmpc_uno_ms_hess.optim)
 
 # TODO: does not work well with MadNLP and MultipleShooting or TrapezoidalCollocation, 
 # figure out why. Current theory: 
@@ -375,6 +381,11 @@ CASE_MPC["Pendulum"]["NonLinMPC"]["Noneconomic"]["MadNLP"]["SingleShooting"] =
         sim!($nmpc_madnlp_ss, $N, $ry; plant=$plant, x_0=$x_0, x̂_0=$x̂_0, progress=false),
         samples=samples, evals=evals, seconds=seconds
     )
+CASE_MPC["Pendulum"]["NonLinMPC"]["Noneconomic"]["Uno"]["MultipleShooting (Hessian)"] = 
+    @benchmarkable(
+        sim!($nmpc_uno_ms_hess, $N, $ry; plant=$plant, x_0=$x_0, x̂_0=$x̂_0, progress=false),
+        samples=samples, evals=evals, seconds=seconds
+    )    
 
 # ----------------- Case study: Pendulum economic --------------------------------
 model2, p = pendulum_model2, pendulum_p2

benchmark/Project.toml

@@ -13,3 +13,4 @@ Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 SparseArrays = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
 SparseConnectivityTracer = "9f842d2f-2579-4b1d-911e-f412cf18a3f5"
 SparseMatrixColorings = "0a514795-09f3-496d-8182-132a7b665d35"
+UnoSolver = "1baa60ac-02f7-4b39-a7a8-2f4f58486b05"

benchmark/benchmarks.jl

@@ -1,6 +1,6 @@
 using BenchmarkTools
 using ModelPredictiveControl, ControlSystemsBase, LinearAlgebra
-using JuMP, OSQP, DAQP, Ipopt, MadNLP
+using JuMP, OSQP, DAQP, Ipopt, MadNLP, UnoSolver
 
 const SUITE = BenchmarkGroup(["ModelPredictiveControl"])
 

src/controller/nonlinmpc.jl

@@ -222,11 +222,11 @@ This controller allocates memory at each time step for the optimization.
    function, see [`DifferentiationInterface` doc](@extref DifferentiationInterface List).
 - `jacobian=default_jacobian(transcription)` : an `AbstractADType` backend for the Jacobian
    of the nonlinear constraints, see `gradient` above for the options (default in Extended Help).
-- `hessian=false` : an `AbstractADType` backend for the Hessian of the Lagrangian, see 
-   `gradient` above for the options. The default `false` skip it and use the quasi-Newton
-    method of `optim`, which is always the case if `oracle=false` (see Extended Help).
-- `oracle=JuMP.solver_name(optim)=="Ipopt"`: use the efficient [`VectorNonlinearOracle`](@extref MathOptInterface MathOptInterface.VectorNonlinearOracle)
-   for the nonlinear constraints (not supported by most optimizers for now).
+- `hessian=false` : an `AbstractADType` backend or `Bool` for the Hessian of the Lagrangian, 
+   see `gradient` above for the options. The default `false` skip it and use the quasi-Newton
+   method of `optim`, which is always the case if `oracle=false` (see Extended Help).
+- `oracle=JuMP.solver_name(optim)=="Ipopt"` : a `Bool` to use the [`VectorNonlinearOracle`](@extref MathOptInterface MathOptInterface.VectorNonlinearOracle)
+   for efficient nonlinear constraints (not supported by most optimizers for now).
 - additional keyword arguments are passed to [`UnscentedKalmanFilter`](@ref) constructor 
   (or [`SteadyKalmanFilter`](@ref), for [`LinModel`](@ref)).
 

src/estimator/mhe/construct.jl

@@ -279,8 +279,8 @@ transcription for now.
 - `hessian=false` : an `AbstractADType` backend for the Hessian of the Lagrangian, see 
    `gradient` above for the options. The default `false` skip it and use the quasi-Newton
     method of `optim`, which is always the case if `oracle=false` (see Extended Help).
-- `oracle=JuMP.solver_name(optim)=="Ipopt"`: use the efficient [`VectorNonlinearOracle`](@extref MathOptInterface MathOptInterface.VectorNonlinearOracle)
-   for the nonlinear constraints (not supported by most optimizers for now).
+- `oracle=JuMP.solver_name(optim)=="Ipopt"` : a `Bool` to use the [`VectorNonlinearOracle`](@extref MathOptInterface MathOptInterface.VectorNonlinearOracle)
+   for efficient nonlinear constraints (not supported by most optimizers for now).
 - `direct=true`: construct with a direct transmission from ``\mathbf{y^m}`` (a.k.a. current
    estimator, in opposition to the delayed/predictor form).