Abstract—Congestion control is a significant challenge in the Transport layer of Mobile Ad Hoc Networks (MANETs). Despite extensive research, the issue remains unresolved. Numerous studies have attempted to address congestion control by integrating solutions with the Medium Access Control (MAC) layer and routing protocols, yet these approaches have encountered various limitations. This article proposes a novel solution focused on congestion control in the Transport layer through buffer management. The proposed solution introduces a Multilevel Queue Management (MQM) algorithm, combined with a packet memory buffer that schedules internal packets based on packet categories and queue assignments. The queue management system categorizes packets into four types: beacon signals, control signals, synchronization signals, and data packets. In the event of network congestion, data packets are redirected by clearing other lower-priority queues to ensure seamless transmission. This queue management system is integrated with the Ad Hoc On-Demand Distance Vector (AODV) protocol, forming the Multilevel Queue Management-Ad Hoc On-Demand Distance Vector (MQM-AODV) protocol. The performance of MQM-AODV (Multi level Management AODV) is evaluated against other algorithms, including Ant Colony Optimization (ACO-AODV), Reinforcement Learning-Ad Hoc On-Demand Distance Vector (RL-AODV), and Categorical Boosting AODV-Ad Hoc On Demand Distance Vector (CATBOOST-AODV). Simulation results, conducted with node counts ranging from 25 to 150, demonstrate that MQM-AODV outperforms these protocols. Specifically, MQM-AODV improves throughput by up to 60.74% over (ACO-AODV) and 36.41% over Reinforcement Learning AODV (RL-AODV), while reducing end-to-end delay by 35.47% compared to ACO-AODV. Although buffer capacity remains similar across protocols, MQM-AODV excels in maximizing network lifetime and minimizing stabilization time, significantly enhancing network reliability. Despite requiring additional scheduling time, the overall performance gains position MQM-AODV as a superior choice for high-demand network environments.


Keywords—Mobile Ad Hoc Networks (MANET), congestion control, multilevel queue management, transport layer,Ad Hoc On-Demand Distance Vector (AODV)