trying_some_RL

Here are some RL projects I tried based off of the ones in the udacity deep reinforcement learning repository. I had to update to the gymnasium package instead of gym, reassess the hyperparameters, and some other tweaks. I also included some sources for myself or others to provide more background. I played around with these agents and modified the code to see the impact of various choices. As some others wrote on deep RL, the hyperparameters can determine whether the agent will or will not solve the problem. This in some instances I found difficult to conceptualize, especially as one such determining parameter could be the random seed in the agent or the environement. While I looked to do bigger reinforcement learning projects in the future, due to the difficulty and the training times, I find these gym environements to be good benchmarks to test whatever agents and perhaps given my current ability and computational power, I should focus on improving times rather than tackling bigger environments.