change RL-Controller initialization process and action space iterable

- no fake builder pattern anymore, moved needed fields into constructor
- add serializeUID
- action space extends iterable interface to simplify looping over all actions (and not returning the actual list)

src/main/java/core/algo/EpisodicLearning.java

@@ -3,12 +3,10 @@ package core.algo;
 import core.DiscreteActionSpace;
 import core.Environment;
 import core.listener.LearningListener;
-import lombok.Getter;
 import lombok.Setter;
 
 public abstract class EpisodicLearning<A extends Enum> extends Learning<A> implements Episodic {
     @Setter
-    @Getter
     protected int currentEpisode;
     protected int episodesToLearn;
     protected volatile int episodePerSecond;

src/main/java/core/algo/Learning.java

@@ -9,14 +9,13 @@ import core.policy.Policy;
 import lombok.Getter;
 import lombok.Setter;
 
-import java.io.Serializable;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
 import java.util.concurrent.CopyOnWriteArrayList;
 
 @Getter
-public abstract class Learning<A extends Enum> implements Serializable {
+public abstract class Learning<A extends Enum>{
     protected Policy<A> policy;
     protected DiscreteActionSpace<A> actionSpace;
     @Setter