fixed link to legacy MDP tutorial (#11)

legacy/rl-tuto/reinforcement_learning_tutorial.ipynb

@@ -87,7 +87,7 @@
     "# Implementation\n",
     "\n",
     "Let's start by defining a type containing all the problem parameters. In Reinforcement Learning, problems are usually framed as Markov decision processes (MDP). Our new problem must then be defined as a subtype of the abstract type `MDP{S,A}` where `S`and `A` are the state type and the action type respectively.\n",
-    "For more information about MDPs please refer to this [MDP tutorial](http://nbviewer.jupyter.org/github/sisl/POMDPs.jl/blob/master/examples/GridWorld.ipynb)."
+    "For more information about MDPs please refer to this [MDP tutorial](../GridWorld.ipynb)."
    ]
   },
   {