Proximal policy optimization with self adaptive penalty function for vehicular resource allocation

Abstract

The vehicle-to-everything (V2X) is a significant technology that improves road safety, travel experience and entertainment services. Resource allocation (RA) in vehicular networks defines the strategic distribution of communication resources like power, time slots and bandwidth among vehicles and infrastructures to provide effective data transmission. However, RA faces challenges in managing limited transmission resources because of network delays and high reception time by frequent changes in network topology and varying user demands through high mobility of vehicles. Therefore, this research proposes a proximal policy optimization with self adaptive penalty function (PPO-SAPF) based RA for vehicular communications. The PPO-SAPF optimizes the RA in dynamic vehicular networks by adjusting the coefficient matrices, which ensures better adaptability to network topology and user demands. The SAPF provides fine-tuning of policy updates, maintaining a better balance between exploration and exploitation, thereby enhancing performance under different network conditions. The PPO-SAPF achieves a less inter-packet reception time of 119 ms for 16 vehicle-to-vehicle (V2V) links in case 3 compared to context-aware RA (CARA). These results demonstrate that the proposed PP-SAPF is suitable for real-time deployment in intelligent transportation systems (ITS) and autonomous vehicles where low latency, reliable connectivity, and adaptive resource management is significant.