diff --git a/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-02c-QuikGraph.ipynb b/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-02c-QuikGraph.ipynb
index 7f1a8f9a71..e42a31c2cf 100644
--- a/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-02c-QuikGraph.ipynb
+++ b/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-02c-QuikGraph.ipynb
@@ -885,7 +885,7 @@
     "// illustrer la CONDITION D'OPTIMALITE de Dijkstra — pour chaque sommet v, son predecesseur\n",
     "// est l'unique voisin u tel que GetDistance(u) + poids(u, v) = GetDistance(v). C'est le seul\n",
     "// site de la serie a rester manuel ; les autres utilisent l'observateur canonique\n",
-    "// VertexPredecessorRecorderObserver (cell 17 de ce notebook, Search-15 cell 14, Search-3\n",
+    "// VertexPredecessorRecorderObserver (cell 17 de ce notebook, Search-02b-NetworkX-Csharp cell 14, Search-3\n",
     "// cell 16, Search-2 cell 38), qui encapsule exactement cette condition.\n",
     "\n",
     "var dijkstra = new DijkstraShortestPathAlgorithm<string, STaggedEdge<string, double>>(\n",
diff --git a/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-3-Informed-Csharp.ipynb b/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-3-Informed-Csharp.ipynb
index 0cc86d4cba..9993af6a51 100644
--- a/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-3-Informed-Csharp.ipynb
+++ b/MyIA.AI.Notebooks/Search/Part1-Foundations/Search-3-Informed-Csharp.ipynb
@@ -924,7 +924,7 @@
     }
    ],
    "source": [
-    "// Installation QuikGraph 2.5.0 (fork KeRNeLith de QuickGraph, voir Search-16 cell 2 pour reference).\n",
+    "// Installation QuikGraph 2.5.0 (fork KeRNeLith de QuickGraph, voir Search-02c-QuikGraph cell 2 pour reference).\n",
     "// Charge la DLL + importe les namespaces necessaires : graphe bidirectionnel + AStar.\n",
     "#r \"nuget: QuikGraph, 2.5.0\"\n",
     "\n",
diff --git a/scripts/notebook_tools/twin_pairs.d/search-3-informed.yaml b/scripts/notebook_tools/twin_pairs.d/search-3-informed.yaml
index 6410aac4d5..eb8d4bd234 100644
--- a/scripts/notebook_tools/twin_pairs.d/search-3-informed.yaml
+++ b/scripts/notebook_tools/twin_pairs.d/search-3-informed.yaml
@@ -60,6 +60,12 @@
       csharp_sha: 0cc86d4cba5b06fb089d75e2154d2b261d5632cf
       content_python_sha: c2795ee971045d3f85e185e6c42fa0633230a70d5ea936b46b060eec49df0775
       content_csharp_sha: 7272c885bb1d9706303987d43dcd12db75839e6c55cb043c2835734a2bc88808
+    - date: "2026-09-03"
+      by: myia-ai-01:CoursIA
+      python_sha: 6a38b61b2c754511513922466155023c4a39ae78
+      csharp_sha: 9993af6a511d618d3a367c7efa61cae62b79799e
+      content_python_sha: c2795ee971045d3f85e185e6c42fa0633230a70d5ea936b46b060eec49df0775
+      content_csharp_sha: ea4e647e62c110746d15c79b2938134fd2a4049f3ab4d6eef0dee81d8dac9392
   known_differences:
     - "Socle pedagogique commun : A* et Greedy Best-First avec differentes heuristiques (distance Manhattan, euclidienne, h=0=degenere en UCS). Arrive en Roumanie + 8-puzzle."
     - "Idiomes framework : Python = `import heapq` (priority queue avec (f, tie_breaker, noeud)). C# = `PriorityQueue<AStarNode, double>` (f-cost = g + h)."