Towards reinterpretation of interaction complexity for load prediction in cloud-based MMORPGs

We present an approach for predicting the load of Massively Multiplayer Online Role-Playing Games (MMORPGs) in a cloud-hosted environment. The load of a MMORPG is determined not only by the number of players, but also by what they do in the virtual world. In our previous work, we have shown that player activity patterns can be categorized, and that each category may be represented by a corresponding mathematical model. In this work we extend the concept of interaction complexity applied for load prediction by Nae et. al. by taking into account MMORPG player behavior categories, and asses the prediction accuracy of the proposed approach.