Here we can observe three experiments with PPO RL algorithm (with different number of timesteps {2000,5000,10000}) seperately (with non-smoothing form) and finally all three together (with smoothing form).

Here we can observe three experiments with DQN RL algorithm (with different number of timesteps {2000,5000,10000}) seperately (with non-smoothing form) and finally all three together (with smoothing form).

Here we can observe three experiments with A2C RL algorithm (with different number of n_steps {16,36}) seperately (with non-smoothing form) and finally both of them (with smoothing form).