Authors: Damijan Novak, Aleš Čep and Domen Verber
In this paper we have addressed the vast fragmentation of the research area of real-time strategy (RTS) games, with the purpose of establishing common ground. We identified four aspects (environmental, hierarchy abstraction, strength and time) that comprise the rich worlds of RTS games. Based on these aspects we created three main classes of RTS game worlds. The classification interval was also defined, which offers the further creation of new (sub)classes if the need arises. We believe a common ground that classes create, should improve the overall field by enabling a comparison of game AI techniques belonging to same class (classes serve as a grouping factor) and can bring positive progress in overall research. In this way, information about the aspects that define the game AI can be exact and the research results can be easily shared and reproduced. Our vision is also to bridge the gap between game AI and classic AI classification techniques.
Keywords: game aspects; game AI classification; game AI; real-time strategy game