Reinforcement Learning is an area of Machine Learning inspired by behaviorist psychology based on the mechanism of learning from rewards. RL does not require prior knowledge and automatically get optimal policy with the help of knowledge obtained by trial-and-error and continuous interaction with the dynamic environment. In complex real world domains implementing RL algorithms is the major practical problem due to the large and continuous space. It can give rise to problems like Curse of Dimensionality, Partial Observability Problem, Credit Structuring Problem, Generalization and Exploration-Exploitation Dilemma. This paper gives an introduction to Reinforcement Learning, discusses its basic model and system structure, and discusses the problems faced while implementing RL algorithms in complex real world domains. At last but not the least this paper briefly describes the techniques which can make the working of RL process easier in the complex domains. It concludes with research scope of RL in complex real world.