add sub_kwargs parameter to LMSolver and solve! for subsolver (#245)

* add sub_kwargs parameter to LMSolver and solve! for subsolver keyword arguments

* add sub_kwargs parameter to LMTRSolver and solve! for passing subsolver keyword arguments and prox evaluations

Author: MohamedLaghdafHABIBOULLAH

src/LMTR_alg.jl

@@ -140,6 +140,7 @@ For advanced usage, first define a solver "LMSolver" to preallocate the memory u
 - `β::T = 1/eps(T)`: TODO
 - `χ =  NormLinf(1)`: norm used to define the trust-region;`
 - `subsolver::S = R2Solver`: subsolver used to solve the subproblem that appears at each iteration.
+- `sub_kwargs::NamedTuple = NamedTuple()`: a named tuple containing the keyword arguments to be sent to the subsolver. The solver will fail if invalid keyword arguments are provided to the subsolver. For example, if the subsolver is `R2Solver`, you can pass `sub_kwargs = (max_iter = 100, σmin = 1e-6,)`.
 
 The algorithm stops either when `√(ξₖ/νₖ) < atol + rtol*√(ξ₀/ν₀) ` or `ξₖ < 0` and `√(-ξₖ/νₖ) < neg_tol` where ξₖ := f(xₖ) + h(xₖ) - φ(sₖ; xₖ) - ψ(sₖ; xₖ), and √(ξₖ/νₖ) is a stationarity measure.
 
@@ -203,6 +204,7 @@ function SolverCore.solve!(
   γ::T = T(3),
   α::T = 1 / eps(T),
   β::T = 1 / eps(T),
+  sub_kwargs::NamedTuple = NamedTuple(),
 ) where {T, G, V}
   reset!(stats)
 
@@ -265,6 +267,7 @@ function SolverCore.solve!(
 
   local ξ1::T
   local ρk::T = zero(T)
+  local prox_evals::Int = 0
 
   residual!(nls, xk, Fk)
   jtprod_residual!(nls, xk, Fk, ∇fk)
@@ -282,6 +285,7 @@ function SolverCore.solve!(
   set_objective!(stats, fk + hk)
   set_solver_specific!(stats, :smooth_obj, fk)
   set_solver_specific!(stats, :nonsmooth_obj, hk)
+  set_solver_specific!(stats, :prox_evals, prox_evals + 1)
 
   φ1 = let Fk = Fk, ∇fk = ∇fk
     d -> dot(Fk, Fk) / 2 + dot(∇fk, d) # ∇fk = Jk^T Fk
@@ -341,22 +345,25 @@ function SolverCore.solve!(
       solve!(
         solver.subsolver,
         solver.subpb,
-        solver.substats,
+        solver.substats;
         x = s,
         atol = stats.iter == 0 ? 1.0e-5 : max(sub_atol, min(1.0e-1, ξ1 / 10)),
         Δk = ∆_effective / 10,
+        sub_kwargs...,
       )
     else
       solve!(
         solver.subsolver,
         solver.subpb,
-        solver.substats,
+        solver.substats;
         x = s,
         atol = stats.iter == 0 ? 1.0e-5 : max(sub_atol, min(1.0e-1, ξ1 / 10)),
         ν = ν,
+        sub_kwargs...,
       )
     end
 
+    prox_evals += solver.substats.iter
     s .= solver.substats.solution
 
     sNorm = χ(s)
@@ -438,6 +445,7 @@ function SolverCore.solve!(
     set_solver_specific!(stats, :nonsmooth_obj, hk)
     set_iter!(stats, stats.iter + 1)
     set_time!(stats, time() - start_time)
+    set_solver_specific!(stats, :prox_evals, prox_evals + 1)
 
     ν = α * Δk / (1 + σmax^2 * (α * Δk + 1))
     @. mν∇fk = -∇fk * ν

src/LM_alg.jl

@@ -140,6 +140,7 @@ For advanced usage, first define a solver "LMSolver" to preallocate the memory u
 - `θ::T = 1/(1 + eps(T)^(1 / 5))`: is the model decrease fraction with respect to the decrease of the Cauchy model;
 - `m_monotone::Int = 1`: monotonicity parameter. By default, LM is monotone but the non-monotone variant will be used if `m_monotone > 1`;
 - `subsolver = R2Solver`: the solver used to solve the subproblems.
+- `sub_kwargs::NamedTuple = NamedTuple()`: a named tuple containing the keyword arguments to be sent to the subsolver. The solver will fail if invalid keyword arguments are provided to the subsolver. For example, if the subsolver is `R2Solver`, you can pass `sub_kwargs = (max_iter = 100, σmin = 1e-6,)`.
 
 The algorithm stops either when `√(ξₖ/νₖ) < atol + rtol*√(ξ₀/ν₀) ` or `ξₖ < 0` and `√(-ξₖ/νₖ) < neg_tol` where ξₖ := f(xₖ) + h(xₖ) - φ(sₖ; xₖ) - ψ(sₖ; xₖ), and √(ξₖ/νₖ) is a stationarity measure.
 
@@ -202,6 +203,7 @@ function SolverCore.solve!(
   η2::T = T(0.9),
   γ::T = T(3),
   θ::T = 1/(1 + eps(T)^(1 / 5)),
+  sub_kwargs::NamedTuple = NamedTuple(),
 ) where {T, V, G}
   reset!(stats)
 
@@ -334,15 +336,16 @@ function SolverCore.solve!(
     solver.subpb.model.σ = σk
     isa(solver.subsolver, R2DHSolver) && (solver.subsolver.D.d[1] = 1/ν)
     if isa(solver.subsolver, R2Solver) #FIXME
-      solve!(solver.subsolver, solver.subpb, solver.substats, x = s, atol = sub_atol, ν = ν)
+      solve!(solver.subsolver, solver.subpb, solver.substats; x = s, atol = sub_atol, ν = ν, sub_kwargs...)
     else
       solve!(
         solver.subsolver,
         solver.subpb,
-        solver.substats,
+        solver.substats;
         x = s,
         atol = sub_atol,
         σk = σk, #FIXME
+        sub_kwargs...,
       )
     end