Project Title: Neural Network Approaches on Performing different tasks on Rogo, a TSP-Based Puzzle Student: Szu-Yi- Wu Course: MSc in Computer Science Abstract: Rogo is a TSP-related puzzle game, which means solving it can be considered as an optimization problem. This project involves two approaches of neural network, solving Rogo puzzles and classification on Rogo puzzles based on their difficulties, which can be used to determine the capability of solving optimization problems using neural network and the capability of classifying the puzzles. Due to the limitation of knowledge and time, the approach of solving puzzle has been discontinued, however, it can still be an interesting work in the future. On the other hand, the classification has provided a decent result, which leads to the thought of it is capable for a neural network to perform a classification based on the difficulties of Rogo. The dissertation will be discussing about the details of this project.