Proposition of a three layers architecture for adaptable artificial intelligence

This paper proposes a concept of artificial intelligence architecture which would be usable in various kind of problems. Artificial intelligences are used in many areas of computer science for decision making tasks. Traditionally each artificial intelligence is designed and programmed to be used within a particular software and for a specific purpose. However, this paper stands as the first step of research in progress whose final objective is to create an artificial intelligence adaptable to all kinds of problems without any change in its source code. The present proposition focuses on the architecture of that artificial intelligence and is introduced in the context of video games. This architecture, composed of three layers, would be re-usable for all types of game.