Large Language Models (LLMs) are increasingly integrated into mobile services over wireless networks to support complex user requests. This trend has led to longer prompts, which improve LLMs' performance but increase data transmission costs and require more processing time, thereby reducing overall system efficiency and negatively impacting user experience. To address these challenges, we propose Joint Prompt and Power Optimization (JPPO), a framework that jointly optimizes prompt compression and wireless transmission power for mobile LLM services. JPPO leverages a Small Language Model (SLM) deployed at edge devices to perform lightweight prompt compression, reducing communication load before transmission to the cloud-based LLM. A Deep Reinforcement Learning (DRL) agent dynamically adjusts both the compression ratio and transmission power based on network conditions and service constraints, aiming to minimize service time while preserving response fidelity. We further extend the framework to JPPO++, which introduces a denoising-inspired compression scheme. This design performs iterative prompt refinement by progressively removing less informative tokens, allowing for more aggressive yet controlled compression. Experimental results show that JPPO++ reduces service time by 17% compared to the no-compression baseline while maintaining output quality. Under compression-prioritized settings, a reduction of up to$16\times$in prompt length can be achieved with an acceptable loss in accuracy. Specifically, JPPO with a$16\times$ratio reduces total service time by approximately 42.3%, and JPPO++ further improves this reduction to 46.5%.

This paper presents a novel approach for wireless Virtual Reality (VR) systems by integrating mobile crowdsensing with Deep Reinforcement Learning (DRL). In modern VR applications, ensuring optimal performance while managing system resources presents challenges including high bandwidth requirements, low Motion-to-Photon delay, and intensive computational demands. We address these challenges by proposing a DRL-based framework that jointly optimizes rendering strategies (local, remote, or collaborative) and user-SBS associations to maximize Quality of Experience while adhering to strict latency constraints. To overcome sparse feedback in complex wireless VR environments, we introduce a two-stage approach combining auction-based IoT-VR pairing with diffusion reasoning-enhanced DRL using the Diffusion Reasoning-based Reward Shaping Scheme (DRESS). Our energy model captures rendering-specific power consumption patterns across different strategies. Simulation results demonstrate that our diffusion-enhanced approach (PPO-RS+DF) achieves superior reward convergence and significantly lower latency compared to baseline methods. The diffusion mechanism effectively propagates sparse reward signals across the state-action space, enabling efficient learning from limited feedback and guiding the learning process toward latency-optimized policies for next-generation wireless VR applications.

Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks, leading to their increasing deployment in wireless networks for a wide variety of user services. However, the growing longer prompt setting highlights the crucial issue of computational resource demands and huge communication load. To address this challenge, we propose Joint Power and Prompt Optimization (JPPO), a framework that combines Small Language Model (SLM)-based prompt compression with wireless power allocation optimization. By deploying SLM at user devices for prompt compression and employing Deep Reinforcement Learning for joint optimization of compression ratio and transmission power, JPPO effectively balances service quality with resource efficiency. Experimental results demonstrate that our framework achieves high service fidelity and low bit error rates while optimizing power usage in wireless LLM services. The system reduces response time by about 17 %, with the improvement varying based on the length of the original prompt.

As a technology intended to reduce cellular network congestion and enhance user service quality, computation offloading in Multi-access Edge Computing (MEC) networks highlights the crucial issue of privacy protection. This paper proposes a novel solution to the computation offloading and privacy protection problem in the MEC network using a Multi-agent Deep Deterministic Policy Gradient (MADDPG) framework. Our approach utilizes game theory to encourage computation offloading by modeling the interaction between cloudlets, Data Center Operator (DCO), and users as an auction game. We formulate the resource allocation and privacy protection as an auction game with multiple bidders and incomplete information and then use MADDPG to find an optimal solution. To ensure privacy protection, we design a Local Differential Privacy (LDP) method in the MADDPG algorithm. Theoretical analysis and simulation results demonstrate the effectiveness of our approach in satisfying differential privacy and converging to an equilibrium. The proposed solution holds significant promise in addressing the computation offloading and privacy protection challenges in MEC networks.

With the burst of mobile data, it is necessary to make use of idle mobile equipment for caching space. Caching in the femto-cells is proposed for reducing transmission latency between the WiFi points and its mobile users (MU) and better user service. In this paper, we firstly take the copyright of the files as the allocation resource and the WiFi points with caching space want to rent these copyrights. We propose a multi-rounds double auction mechanism for this problem and take the popularity parameter of the files as the quality weight. This game can help multiple content providers (CP) lease copyrights of the files to multiple WiFi points effectively. Different from traditional double auctions, it will take the failed buyers and sellers into consideration and they are allowed to change their requests in the next auction process. This mechanism can largely improve efficiency of the game and is budget balanced. We also prove that the allocation process is monotone and with the set of the critical payment rule, we prove the truthfulness of the mechanism. Additionally, we prove that the mechanism can form a conditional equilibrium. Simulation results verify the effectiveness of the proposed mechanism and compare with the traditional one-round double auction.