With the evolution of artificial intelligence-generated content (AIGC) techniques and the development of space-airground integrated networks (SAGIN), there will be a growing opportunity to enhance more users' mobile experience with customized AIGC applications. This is made possible through the use of parameter-efficient fine-tuning (PEFT) training alongside mobile edge computing. In this paper, we formulate the optimization problem of maximizing the parameter training efficiency of the SAGIN system over wireless networks under limited resource constraints. We propose the Parameter training efficiency Aware Resource Allocation (PARA) technique to jointly optimize user association, data offloading, and communication and computational resource allocation. Solid proofs are presented to solve this difficult sum of ratios problem based on quadratically constrained quadratic programming (QCQP), semidefinite programming (SDP), graph theory, and fractional programming (FP) techniques. Our proposed PARA technique is effective in finding a stationary point of this non-convex problem. The simulation results demonstrate that the proposed PARA method outperforms other baselines.

No abstract

With the vigorous development of information and communication technology, mobile internet has undergone tremendous changes. How to achieve global coverage of the network has become the primary problem to be solved. GEO satellites and LEO satellites, as important components of the satellite–ground network, can offer service for hotspots or distant regions where ground-based base stations’ coverage is limited. Therefore, we build a satellite–ground network model, which transforms the satellite–ground network resource allocation problem into a matching issue between GEO satellites, LEO satellites, and users. A GEO satellite provides data backhaul for users, and a LEO satellite provides data transmission services according to users’ requests. It is important to consider the relationships between all entities and establish a distributed scheme, so we propose a three-sided cyclic matching algorithm. It is confirmed by a large number of simulation experiments that the method suggested in this research is better than the conventional algorithm in terms of average delay, satellite revenue, and number of users served.

No abstract

No abstract