Improve efficiency of set_normalized_rhs! with exact Hessians #377
Description
Similarly to jump-dev/Ipopt.jl#515, calling set_normalized_rhs! to update the constant vector of linear constraints before re-solving is not efficient, since it invalidates the inner model thus requiring an expensive initialize calls.
Here's a minimum reproducible example with ModelPredictiveControl.jl:
using ModelPredictiveControl, JuMP
using KNITRO
N = 35 # number of JuMP.optimize! calls
function f!(ẋ, x, u, _ , p)
g, L, K, m = p # [m/s²], [m], [kg/s], [kg]
θ, ω = x[1], x[2] # [rad], [rad/s]
τ = u[1] # [Nm]
ẋ[1] = ω
ẋ[2] = -g/L*sin(θ) - K/m*ω + τ/m/L^2
return nothing
end
h!(y, x, _ , _ ) = (y[1] = 180/π*x[1]; nothing) # [°]
p = [9.8, 0.4, 1.2, 0.3]
nu, nx, ny, Ts = 1, 2, 1, 0.1
model = NonLinModel(f!, h!, Ts, nu, nx, ny; p)
p_plant = copy(p); p_plant[3] = p[3]*1.25
plant = NonLinModel(f!, h!, Ts, nu, nx, ny; p=p_plant)
Hp, Hc, Mwt, Nwt = 20, 2, [0.5], [2.5]
α=0.01; σQ=[0.1, 1.0]; σR=[5.0]; nint_u=[1]; σQint_u=[0.1]
σQint_ym = zeros(0)
umin, umax = [-1.5], [+1.5]
transcription = MultipleShooting()
optim = Model(KNITRO.Optimizer)
oracle = true
hessian = true
nmpc = NonLinMPC(model;
Hp, Hc, Mwt, Nwt, Cwt=Inf, transcription, oracle, hessian, optim,
α, σQ, σR, nint_u, σQint_u, σQint_ym
)
nmpc = setconstraint!(nmpc; umin, umax)
unset_time_limit_sec(nmpc.optim)
sim!(nmpc, N, [180.0]; plant=plant, x_0=[0, 0], x̂_0=[0, 0, 0])
@time sim!(nmpc, N, [180.0]; plant=plant, x_0=[0, 0], x̂_0=[0, 0, 0])
@profview sim!(nmpc, N, [180.0]; plant=plant, x_0=[0, 0], x̂_0=[0, 0, 0])producing this profile:
The initialize portion would completely disappear of the flame graph if the model was not invalidated. A solution similar to jump-dev/Ipopt.jl#516 could be re-implemented here.
Thanks!