Abstract - In recent years, several sequential decision problems in robotics such as motion planning, hand minpulation etc.., are solved using reinforcement learning. In this project we would like to study and analyse reinforcement learning algorithms in continuous action and state space settings, which are common in robotics. We would like to understand (description of how the algorithms work, theoretical analysis and convergence properties) both model based algorithms such as iLQR, DDP, iLQG and model free policy gradient algorithms like actor-critic, Reinforce, pi^2. Further, we aim to implement some of these algorithms (atleast one model based and one model free) and experiment with them on two different environments/robot systems (ex : the inverted cartpole). This would allow comparison of these algorithms on a common platform and draw conclusions on suitability of these algorithms for these settings.