Wireless Metaverse Behavior Models and Optimization Based on Bandwagon Effects

Users’ behaviors in wireless metaverse networks are usually affected by the surrounding people and limited network resources. How to allocate network resources in a human-centric way remains an open problem in wireless metaverse scenarios. To capture the psychological influence of bandwagon effects on users’ behaviors, we first propose bandwagon effect-based metaverse behavior metrics, including the metaverse bandwagon threshold and metaverse bandwagon probability, based on the multi-dimensional contract theory. The bandwagon effect-based metaverse behavior metrics are used to quantify the number of service adopters, which consider both users’ behaviors and resource allocation strategies. Moreover, the metaverse behavior utility is derived for wireless metaverse networks based on the multi-attribute utility theory. To solve the metaverse behavior utility maximization problem, a bandwagon effect optimal transport-based (BEOT) algorithm is proposed to optimize the resource allocation strategies considering users’ behavior characteristics. Compared with the maximum metaverse behavior utility of virtual reality tracking-based resource allocation (VRT), soft actor-critic with graph convolutional networks (SAC-GCN) and the deep Q-learning-based (DQL) algorithms, simulation results show that the maximum metaverse behavior utility of proposed BEOT algorithm is improved by 22.55%, 12.36% and 18.21%, respectively.