Multi-type service placement for joint transmission and processing in mobile augmented reality

Abstract

Augmented Reality (AR) technology can achieve real-time responsiveness and efficient processing by leveraging Edge Computing. Deploying mobile AR applications on edge servers significantly reduces service latency and enhances system quality of service (QoS). However, existing studies primarily focus on service placement on edge servers, while often neglecting the critical impact of network access points on service latency. In this study, we propose a joint transmission and processing multi-type service placement framework. This framework first determines the user’s access point and then identifies the optimal location for deploying the application serving that user, aiming to minimize the total system service latency. The framework faces two key challenges: user mobility and the diversity of service demands. To address these challenges, this paper designs a heuristic approach based on Block Coordinate Descent (HBCD). This algorithm dynamically selects the optimal access point and edge server for users during each epoch, effectively reducing total system service latency. Theoretical analysis demonstrates that the discrepancy in performance between the proposed algorithm and the optimal solution is tightly bounded. Furthermore, extensive simulations driven by real-world data demonstrate that the HBCD algorithm reaches near-optimal performance and markedly surpasses existing methods.