RPG fallback (#518)

* Add warm start to CRPG backtracking
* Bump changelog
* Fix some math rendering in docs
* Bump changelog
* Add some references
* Update changelog
* Update Changelog.md
* Runic: Update test/solvers/test_proximal_gradient_method.jl
---------

Co-authored-by: Ronny Bergmann <git@ronnybergmann.net>

Author: hajg-ijk

Changelog.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 
 * `CubicBracketingLinesearch` step size
+* fallback in `proximal_gradient_plan`to use the norm of the inverse retraction if the distance is not available.
 
 ## [0.5.23] September 14, 2025
 

Project.toml

@@ -1,7 +1,7 @@
 name = "Manopt"
 uuid = "0fc0a36d-df90-57f3-8f93-d78a9fc72bb5"
 authors = ["Ronny Bergmann <manopt@ronnybergmann.net>"]
-version = "0.5.23"
+version = "0.5.24"
 
 [deps]
 ColorSchemes = "35d6a980-a343-548e-a6ea-1d62b119f2f4"

src/plans/proximal_gradient_plan.jl

@@ -508,8 +508,7 @@ function (s::ProximalGradientMethodBacktrackingStepsize)(
 
         # Compute log_p(candidate_point) and its squared norm for the conditions
         log_p_q = inverse_retract(M, p, candidate_point, st.inverse_retraction_method)
-        squared_distance = distance(M, p, candidate_point)^2
-
+        squared_distance = norm(M, p, log_p_q)^2
         if s.strategy === :nonconvex
             # Nonconvex descent condition
             if get_cost(mp, p) - get_cost(mp, candidate_point) >=

test/solvers/test_proximal_gradient_method.jl

@@ -223,7 +223,14 @@ using Manopt, Manifolds, Test, ManifoldDiff, ManoptTestSuite
         prox_h(M, λ, p) = p
         p0 = p1
         pbm_s = proximal_gradient_method(
-            M, f, g, grad_g; prox_nonsmooth = prox_h, return_state = true
+            M, f, g, grad_g;
+            prox_nonsmooth = prox_h,
+            inverse_retraction_method = ProjectionInverseRetraction(),
+            stepsize = ProximalGradientMethodBacktracking(;
+                initial_stepsize = 1.0,
+                strategy = :convex
+            ),
+            return_state = true
         )
         @test startswith(
             repr(pbm_s), "# Solver state for `Manopt.jl`s Proximal Gradient Method\n"