Multi-Service Edge Computing Management With Multi-Stage Coalition Game Task Offloading

The advent of 5G-enabled edge servers presents an opportunity to distribute computational tasks to the network edge. This approach helps alleviate the strain on limited central network resources caused by the rapid growth in the number of mobile devices and computation-intensive services. Moreover, it leads to reduced end-to-end delays for users. In this paper, we investigate resource allocation optimization in a dynamic multi-service system, where each service provider (SP) serves geographically dispersed service subscribers (SSs). Each SP can offload tasks to multiple edge servers, while each SS can freely switch between SPs offering homogeneous services. We propose the Multi-Stage Coalition Game Task Offloading (MSCGTO) framework, accommodating scalability, resource heterogeneity, and dynamic conditions. This framework encompasses two distributed algorithms to jointly maximize SP profit and minimize SS end-to-end delay, addressing cost-benefit considerations and user latency acceptance. We conduct extensive simulations and practical experiments with real-world services including augmented reality (AR), online gaming, and live video streaming applications, performed in a controlled testbed environment. The results of our experiments demonstrate that the proposed algorithms yield a 25% increase in system utility considering both the profit of SPs and the end-to-end delay of SSs when compared to existing approaches.