Modified the documentation of the Heuristic method (default action = previous action) (#4174)

* Modifying the documentation to explain that Heuristic method default action will be the previous action decided by the heuristic. Changing this behavior would be a breking change.

* Rephrase the working of the documentation of the default action of the Heuristic method

* Forgot an import

Author: vincentpierre

Project/Assets/ML-Agents/Examples/Soccer/Scripts/AgentSoccer.cs

@@ -166,6 +166,7 @@ public override void OnActionReceived(float[] vectorAction)
 
     public override void Heuristic(float[] actionsOut)
     {
+        Array.Clear(actionsOut, 0, actionsOut.Length);
         //forward
         if (Input.GetKey(KeyCode.W))
         {

Project/Assets/ML-Agents/Examples/WallJump/Scripts/WallJumpAgent.cs

@@ -242,6 +242,7 @@ public override void OnActionReceived(float[] vectorAction)
 
     public override void Heuristic(float[] actionsOut)
     {
+        System.Array.Clear(actionsOut, 0, actionsOut.Length);
         if (Input.GetKey(KeyCode.D))
         {
             actionsOut[1] = 2f;

com.unity.ml-agents/Runtime/Agent.cs

@@ -761,6 +761,8 @@ public virtual void Initialize() {}
         ///
         /// Your heuristic implementation can use any decision making logic you specify. Assign decision
         /// values to the float[] array, <paramref name="actionsOut"/>, passed to your function as a parameter.
+        /// The same array will be reused between steps. It is up to the user to initialize
+        /// the values on each call, for example by calling `Array.Clear(actionsOut, 0, actionsOut.Length);`.
         /// Add values to the array at the same indexes as they are used in your
         /// <seealso cref="OnActionReceived(float[])"/> function, which receives this array and
         /// implements the corresponding agent behavior. See [Actions] for more information