Adaptive Multi-Discriminator WGAN Framework for Resource-Constrained Internet of Vehicles Using Reinforcement Learning and Game Theory

Farhoud Jafari Kaleibar; Amr M. Zaki; Marin Litoiu

cs.LGOct 7, 2026

Adaptive Multi-Discriminator WGAN Framework for Resource-Constrained Internet of Vehicles Using Reinforcement Learning and Game Theory

Authors: Farhoud Jafari Kaleibar, Amr M. Zaki, Marin Litoiu

Organizations: Department of Electrical and Computer Engineering, York University, Toronto, ON, Canada

Abstract

Managing machine learning workloads as a network service introduces a resource-orchestration problem distinct from conventional model training; which nodes should be allocated to a task, how communication and computation budgets should be divided among them, and how service quality should be sustained as connectivity and node availability change with mobility. Deploying Generative Adversarial Networks (GANs) in Internet of Vehicles (IoV) environments is a demanding instance of this problem; resource constraints, dynamic network topologies, and competing optimization objectives mean that traditional GAN architectures cannot simultaneously achieve high accuracy, efficient resource use, low delay, and low communication overhead. This paper introduces an adaptive multi-discriminator Wasserstein GAN (MD-WGAN) framework that integrates reinforcement learning with game-theoretic coordination to address these challenges jointly. In our framework, roadside units host generators paired with Deep Q-Network (DQN) agents that select discriminator subsets and manage distributed training across mobile vehicular nodes, while a game-theoretic coordination step allocates training epochs between generators and discriminators. A unified optimization objective ties adversarial learning quality to resource efficiency, communication overhead, and latency under vehicular constraints, allowing the framework to continuously adapt its training behavior as network conditions change. Evaluation on real-world NGSIM trajectory data shows that the framework attains prediction accuracy comparable to state-of-the-art GAN baselines - the lowest RMSE (1.029) and MAE (0.894) among all evaluated methods - while markedly improving resource efficiency: average CPU utilization is reduced by roughly 28% and mean memory usage by roughly 6%, at competitive communication overhead and latency.

Figures & tables

Fig. 1: Architectural view of the proposed framework

Symbol	Description
$G=\{1,\dots,N_{G}\}$	Set of generator nodes (RSUs).
$D=\{1,\dots,K\}$	Set of discriminators (vehicles).
$V=G\cup D$	All nodes in the system.
$g_{i},D_{j}$	Indices for generator and discriminator.
$D_{g_{i}}$	Discriminator subset assigned to $g_{i}$ .
$z\!\sim\!p_{z}$ , $x\!\sim\!P_{r}$ , $\tilde{x}\!\sim\!P_{g}$	Latent, real, and generated samples.

TABLE I: List of Symbols and Variables

Fig. 2: Workflow of proposed approach methodology

Parameter	Value
Simulation Area	1500 m $\times$ 1500 m
Simulation Time	up to 150 seconds
Vehicle Speed Range	15–60 m/s
Mobility Model	SUMO-based Urban Traffic
Communication Protocol	DSRC
Wireless Standard	IEEE 802.11p

TABLE II: Simulation Parameters

Parameter	Value
State space dimension $\|\mathcal{S}\|$	7
Network architecture	FC $[128,128,\|\mathcal{A}\|]$ , ReLU
Discount factor $\gamma$	0.99
Initial exploration $\epsilon_{0}$	1.0
Minimum exploration $\epsilon_{\min}$	0.01
Exploration decay $\epsilon_{\text{decay}}$	0.995

TABLE III: RL Agent (DQN) Hyperparameters

Parameter	Value
Discriminator effectiveness rate $\kappa_{D}$	0.05
Generator effectiveness rate $\kappa_{G}$	0.10
Per-epoch cost $c_{D}$	0.3
Per-epoch cost $c_{G}$	0.8
Discriminator cost weight $\lambda_{D}$	0.05
Generator cost weight $\lambda_{G}$	0.05

TABLE IV: Game-Theoretic Coordination Parameters

Fig. 3: Training cost progression of the proposed Adaptive MD-WGAN.

Fig. 4: Performance comparison of the proposed approach in terms of RMSE and MAE by increasing the training iteration

Fig. 5: Performance Comparison of the proposed approach in terms of CPU and memory usage during simulation time

Fig. 6: Comparison of the proposed method in terms of Network Overhead

Fig. 7: Performance Comparison of the proposed approach in terms of Training Latency

Aspect	Complexity	Implications
Computational	$\mathcal{O}(E_{\text{epochs}}\cdot d\cdot B)$	Scales with training parameters but is ”parallelizable” across nodes. Adding more nodes does not increase this complexity.
Communication	$\mathcal{O}(S_{\text{sync}}\cdot N_{G}\cdot K_{S}\cdot d)$	The communication overhead scales linearly with the number of generator nodes ( $N_{G}$ ) and its related discriminators ( $K_{S}$ ). The framework’s success hinges on adaptive strategies (such as Intelligent Communication Scheduling) to minimize the number of synchronization steps ( $S_{\text{sync}}$ ) and the number of active discriminators ( $\|K\|$ ) per generator.

TABLE V: Summary of Scalability Analysis

Fig. 8: Computational scalability of the proposed approach

Fig. 9: Communication scalability of the proposed approach

Fig. 10: Comparative performance of the proposed approach for different apps in terms of varying metrics

Explore similar work

Jul 26, 2026cs.NI

GNN-based Multi-Agent Control of Traffic Shockwaves in Sparse Vehicular Ad-hoc Networks

Traffic shockwaves are stop-and-go waves that propagate upstream through the streams of vehicles and are one of the major causes of traffic congestion, fuel inefficiency, and increased accident rates in modern transportation systems. Although Connected and Autonomous Vehicles (CAVs) offer a promising opportunity to mitigate such shockwaves, most existing control strategies rely on global traffic state information, making them impractical for early-stage deployment of Vehicular Ad-hoc Networks (VANETs). In this paper, we propose a decentralized Multi-Agent Reinforcement Learning (MARL) framework that integrates a Graph Neural Network (GNN) to enhance the control architecture of connected and autonomous vehicles. The proposed approach enables vehicles to learn cooperative control policies using locally available information and interaction with neighboring vehicles. The effectiveness of the proposed scheme is evaluated using a scalable simulation environment under realistic highway traffic conditions. Simulation results show that the proposed GNN-based MARL framework can reduce the propagation of traffic shockwaves by up to 80%, even when only 10% of the vehicles are connected.

Prachi Nandi, Madhuri Malakar, Sonakshi Satpathy +1

Microsoft · Dept. of Computer Science and Engineering GITAM University Bengaluru, Karnataka · NatWest Group +1

Multi-Agent Control Decentralized MARL

Ask PDF

May 6, 2026cs.NI

Joint Optimization of Trajectory Control, Resource Allocation, and Task Offloading for Multi-UAV-Assisted IoV

This paper investigates a multi-Unmanned Aerial Vehicle (UAV) joint base station-assisted Internet of Vehicles (IoV) task offloading system in dense urban environments. To minimize system delay and energy consumption under strict coupling constraints, the complex non-convex optimization problem is decoupled into a hierarchical execution framework. First, a sequential distributed optimization algorithm based on Second-Order Cone Programming (SOCP) is proposed to optimize the 3D flight trajectory of each UAV, ensuring adaptive network coverage. Second, a novel hybrid resource scheduling paradigm synergizing Deep Reinforcement Learning (DRL) and Large Language Models (LLMs) is developed. Within this framework, the DRL agent dictates the initial resource allocation, while the LLM acts as a semantic macro-scheduler to rectify long-tail allocation imbalances for failed and surplus tasks. Crucially, a reward decoupling mechanism is introduced to isolate DRL training from external LLM interventions, thereby ensuring policy convergence. Finally, the task offloading ratios are precisely determined via Linear Programming (LP) within an alternating optimization loop. Simulation results demonstrate that the proposed method significantly outperforms traditional multi-agent reinforcement learning baselines in terms of task success rate and system efficiency.

Maoxin Ji, Qiong Wu, Pingyi Fan +4

School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China · School of Information Engineering, Jiangxi Provincial Key Laboratory of Advanced Signal Processing and Intelligent Communications, Nanchang University, Nanchang 330031, China · Department of Electronic Engineering, State Key laboratory of Space Network and Communications, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing 100084, China +4

Trajectory Optimization Intelligent Transportation Systems UAV Trajectory Optimization Wireless Resource Allocation LLM Request Scheduling

Ask PDF

Sep 3, 2026cs.NI

From Prior-Guided Heuristics to Deployable Agents: Accelerating Demonstration-Driven Reinforcement Learning for Deadline-Constrained Network Control

Timely delivery of delay-sensitive information over dynamic, heterogeneous networks is essential for NextG interactive applications, yet providing strict End-to-End (E2E) peak latency guarantees remains an open challenge. Two obstacles limit the adoption of learning-based network control in this setting: traditional volume-based routing metrics, while highly effective for general traffic management, are not designed to capture traffic urgency; and Deep Reinforcement Learning (DRL) controllers trained from scratch suffer from sample inefficiency, long training times, and early-stage exploration volatility. This paper introduces a deployment-focused network control framework that addresses both obstacles. First, we present Effective Congestion (EC), a deadline-aware metric family that quantifies interface congestion by packet urgency and proactively filters non-viable traffic, coupled with a Uniform Path Grouping (UPG) distribution heuristic promoting robust load-balancing; the resulting policies are embedded into Multi-Agent Deep Reinforcement Learning Effective Congestion (

p^*

) (MADRL EC (

p^*

)), a hybrid architecture combining a distributed scheduler with a centralized RL-based router. Second, we introduce a unified training objective that generalizes existing policy-learning paradigms---behavioral cloning, offline Reinforcement Learning (RL), online RL, and offline-to-online schemes---as special cases, combining a live-reward term, a pre-collected-reward term, and a policy-imitation term. From this objective, we derive the Model-Guided Annealed Reinforcement Learning (MGA-RL) protocol, instantiated on a Deep Deterministic Policy Gradient (DDPG) backbone: a deployment-oriented, demonstration-driven training approach that generalizes conventional Offline-to-Online (O2O) schemes, in which trajectories from a lightweight [...]

Vincenzo Norman Vitale, Mohammad Solki, Antonia Maria Tulino +2

DIETI, University of Naples Federico II, Via Claudio 21, Naples, 80125, Italy · New York University, 6 MetroTech Center, Brooklyn, New York, 11201, USA · University of Southern California, 3740 McClintock Ave., Los Angeles, 90089, California, USA +2

Learning from Demonstration Multi-Agent Reinforcement Learning Offline-To-Online RL

Ask PDF