doc: adding postprocessing in default sparse backend

Author: franckgaga

src/controller/nonlinmpc.jl

@@ -291,13 +291,16 @@ NonLinMPC controller with a sample time Ts = 10.0 s:
     AutoSparse(
         AutoForwardDiff(); 
         sparsity_detector  = TracerSparsityDetector(), 
-        coloring_algorithm = GreedyColoringAlgorithm((
+        coloring_algorithm = GreedyColoringAlgorithm(
+            (
             NaturalOrder(),
             LargestFirst(),
             SmallestLast(),
             IncidenceDegree(),
             DynamicLargestFirst()
-        ))
+            ), 
+        postprocessing = true
+        )
     )
     ```
     that is, it will test many coloring orders at preparation and keep the best. This is

src/estimator/mhe/construct.jl

@@ -382,13 +382,16 @@ MovingHorizonEstimator estimator with a sample time Ts = 10.0 s:
     AutoSparse(
         AutoForwardDiff(); 
         sparsity_detector  = TracerSparsityDetector(), 
-        coloring_algorithm = GreedyColoringAlgorithm((
+        coloring_algorithm = GreedyColoringAlgorithm(
+            (
             NaturalOrder(),
             LargestFirst(),
             SmallestLast(),
             IncidenceDegree(),
             DynamicLargestFirst()
-        ))
+            ), 
+        postprocessing = true
+        )
     )
     ```
     that is, it will test many coloring orders at preparation and keep the best.