Wonders of Seabed: Difficulty Evaluation of Management Games Using Neural Network

摘要

In management games, players enjoy developing the virtual objects, such as avatars, farms, villages, cities, resources, etc. Usually, the management games do not have a time limit to play. Players know little about how much time that they need to devote in order to quickly build up the virtual objects. This paper studies about the difficulty level of the management games. That is, under a limit amount of time, how far a player can develop the objects. We propose to adopt a neural network to evaluate the difficulty of such kind of games. There is a diverse set of features supported by management games. Thus, we have developed a manageable management game called Wonders of Seabed. Our game is a 3D game and the game story happens at the seabed. Players need to develop a city by managing different resources. Our method can adjust the game difficulty for the players.