Exploiting virtual objects' attributes and avatar's behavior in DVEs partitioning

Abstract

Partitioning constitutes one of the most critical challenges a distributed virtual environment needs to handle and is related to the efficient assignment of the existing entities of the virtual world to the available resources of the system. The efficient partitioning arises from the need of high consistency maintenance among participating users' view, as virtual reality applications are characterized by the high degree of realism they aim to achieve. To the direction of partitioning algorithms, a lot of work has been done, which has been more intense due to the evolution of network technologies and the familiarization of users with virtual reality systems. This paper presents a partitioning approach, which is based on the degree of interaction of each virtual object and the behavior of the participating users, in terms of moving trends, which is usually directed by the actions they can perform on the surrounding objects. In particular, the partitioning takes place when the virtual world contains only objects, in the sense that no avatars are yet connected and objects' attributes are handled as indicators for avatars' future behavior in the system. This prediction of avatars' behavior based on objects' attributes defines the initial partitioning of the virtual space.