Offline Reinforcement Learning is a relatively new area of research within the field of reinforcement learning, and it has been shown to be able to offer significant advantages over its online counterparts. This research project aimed to provide more context on the topic of offline reinforcement learning by discussing some of the benefits it can bring over traditional algorithms, as well as the challenges that are faced during implementation and deployment. Several notable offline reinforcement learning models were also examined in detail, and were used to conduct a comparative analysis. This analysis was performed in a controlled setting and by testing the algorithms on a set of different benchmark tasks in order to provide a standardized evaluation of their performances. Further experiments and parameter-tuning were also conducted to provide additional insights into these models.