Hota, LopamudraKumar, ArunChong, Peter Han Joo2026-04-072026-04-072026-03-24IEEE Internet of Things Journal, ISSN: 2372-2541 (Print); 2327-4662 (Online), Institute of Electrical and Electronics Engineers (IEEE), 1-1. doi: 10.1109/jiot.2026.36772512372-25412327-4662http://hdl.handle.net/10292/20877In Vehicular Ad-hoc Networks (VANETs) and major transportation systems, efficient communication protocol is vital for timely data transmission to vehicles. The dense vehicular network poses challenges to efficient channel-sharing. For the proper utilization of the available bandwidth, optimization of channel mechanisms is crucial. The proposed approach enables vehicles to dynamically tune their Contention Windows (CWs) using locally observable MAC-layer information, with the objective of jointly maximizing throughput, minimizing delay, and maintaining fair channel access. Comprehensive simulations and analysis show notable improvement in the overall network efficiency in terms of throughput, collision, and delay. The adaptiveness of the proposed algorithm guarantees flexibility to changing traffic conditions and is well-suited to the evolving Intelligent Transportation Systems (ITS). With an emphasis on high throughput, low latency, and fair channel allocation, the proposed model contributes to the advanced communication protocols for VANETs. The proposed model also highlights the significance of intelligent adaptive techniques in obtaining enhanced network performance.This is the Author's Accepted Manuscript of an article published in the IEEE Internet of Things Journal © Copyright 2026 IEEE. The Version of Record can be found at DOI: 10.1109/jiot.2026.367725146 Information and Computing Sciences4006 Communications Engineering40 Engineering0805 Distributed Computing1005 Communications TechnologiesVANETMACDRLMulti-AgentContention WindowAdaptiveChannelActor-CriticJournal ArticleOpenAccess10.1109/jiot.2026.3677251