This paper investigates the vehicle to vehicle (V2V) task offloading problem in a vehicular network with multiple radio access technologies (RATs). The problem is formulated as a mixed integer nonlinear programming problem (MINP) with an objective to minimize the average offloading delay of all offloading tasks in the network by optimizing a set of offloading decisions subject to delay and computing resource constraints. To solve the formulated problem, a log-sum-exp approximation is used to transfer the MINP problem to a combinatorial optimization problem to obtain a probability distribution of all possible offloading decisions. Based on the probability distribution, a Markov-chain based method is used to obtain the state transition probabilities between different offloading decisions. Based on the state transition probabilities, a V2V task offloading decision (V2V-TOD) algorithm is proposed to make offloading decisions via using a learning process consisting of three stages: SV and RAT selection, sub-channel selection, computing resource allocation. Simulation results show that the proposed V2V-TOD algorithm can efficiently reduce the average offloading delay as compared with three benchmark algorithms.

In this paper, we investigate the stability and effectiveness of intelligent reflecting surface (IRS) aided systems in the context of mobile multi-users and time-varying channel status. Different from the previous researches in the IRS-aided communication mostly based on one or more independent channel realization, we consider dynamic channel status varying with the mobility of users. Specifically, a dynamic problem as maximizing the time-average rate of all users is formulated. A fractional programming method based on Lagrangian dual theory is proposed as a solution. Simulation results demonstrate that the IRS can be more efficient than amplified forward (AF) relay in adapting the dynamically changing channels stably.

Abstract The hybrid visible light communication (VLC)/radio frequency (RF) communications are investigated with the aid of reconfigurable intelligent surfaces (RISs) in dynamic wireless networks, where the RIS access selection processes of VLC/RF users are updated depending on the channel quality dynamically. Specifically, a dynamic optimization problem due to the mobility of users and time‐varying selection strategy is formulated. Under the constraints of the average minimum rate for VLC/RF users and the maximum transmit power constraints for VLC/RF access points (APs), the target is to minimize the average long‐term power consumption, by jointly considering the active beamforming at APs and the passive beamforming at RISs. Based on the Lyapunov optimization framework and the drift‐plus‐penalty (DPP) algorithm, the original optimization problem is transformed into corresponding short‐term problems at each frame. Furthermore, the closed form solutions with active‐passive beamforming are derived using the fractional programming method based on the Lagrangian dual theory. Finally, numerical results demonstrate the convergence and effectiveness of the proposed optimization algorithm.

Abstract The design of dynamic wireless networks assisted by reconfigurable intelligent surfaces (RIS) mounted on aerial platforms (RIS‐APs) is conceived, where the connection status among users and RIS‐APs are selected according to the average channel quality dynamically and timely. Taking into account the time‐varying selection status and the mobility of users, we construct a long‐term dynamic process. The goal is to minimize the time‐averaged power consumption under the requirements of the time‐averaged minimum rate for users as well as the constraint of the maximum transmit power for the base station (BS), via jointly optimizing the active beamforming at the BS and passive beamforming at RIS‐APs. With the aid of Lyapunov concept‐based drift‐plus‐penalty (DPP) algorithm, the long‐term optimization problem is transformed into short‐term sub‐problems related to each other at each frame. Subsequently, the fractional programming method based on Lagrangian dual theory is applied to derive the solutions for active‐passive beamforming in a closed form. Finally, simulation results validate the convergence and effectiveness of the proposed algorithm.

In this paper, we conceive the design of dynamic wireless networks assisted by multiple intelligent reflecting surfaces (IRSs), where the connection states between users and IRSs are capable of being updated timely. Taking into account the time-varying states of the system, we further construct a long-term dynamic process. Our goal is to maximize the time average sum-rate of the dynamic system under the time average rate and power constraints of users, via jointly optimizing the power allocation at users and the reflecting coefficients at IRSs. With the aid of Lyapunov concept-based drift-plus-penalty (DPP) algorithm, the long-term optimization problem is formulated as an infinite-horizon time-average one. Subsequently, the fractional programming method based on Lagrangian dual transform is applied to optimize power allocation and reflecting coefficients in an iterative manner, and the closed-form solutions of power and reflecting coefficients can be obtained at each iteration. Finally, simulation results demonstrate the convergence and effectiveness of the proposed algorithm. Further performance comparisons indicate that the proposed algorithm can maintain a balance between supply and demand for resource allocation and improve the fairness of users.

This paper investigates the long-term effectiveness and stability of an integrated unmanned aerial vehicles (UAV)-intelligent reflecting surface (IRS) relaying dynamic system in the context of time-varying system states. Consequently, a dynamic optimization problem is constructed to minimize the frame-average transmit power by joint active beamforming at the base station (BS) and passive beamforming at the IRS under frame-average rate constraints. The original problem as an infinite-horizon time-average one can be solved by introducing the drift-plus-penalty (DPP) algorithm and then the optimal active beamforming and passive beamforming can be obtained in an iterative manner. Simulation results demonstrate the theoretical analysis and assess the performance of the dynamic system.