Included random initial positions and random perturbations to cartpole problem (month 23)

This month we focused on the two following goals:

• Make cartpole learning stable enough making use of DQN, adaptative learning and replay buffers
• Investigate about reward policies alternatives and some ways to mitigate catastrophic forgetting

• Include monitorization to cartpole to know how solid is our solution (TODO include this monitorization as a common library in RL-Studio)

• you can see more details about the projec status in cartpole project post