Simulating Human-like Navigation in Urban Transportation Environments with Multi-agent Deep (Inverse) Reinforcement Learning

Abstract

This paper focuses on an approach that uses two methodologies: Maximum Entropy Inverse Reinforcement Learning and Multi-agent Deep reinforcement learning. It stress on each method in a sufficiently profound manner. The nature of data is also specified and it is also displayed how they prepared it to suit their purpose. The use of Keras as one of their tools to construct modest deep neural networks. For results, they found the relation between the outcomes of each method with proper explanation. They also stated the direction in which the approach can be led forward in the future.