SIGNATURE WORK
CONFERENCE & EXHIBITION 2024

With the growing public concern about traffic congestion and traffic safety, intelligent transportation systems (ITS) driven by data and assisted by AI have attracted much attention. As a computing offloading mode based on the Internet of things (IoTs), mobile edge computing (MEC) effectively enhances the efficiency of information exchange between mobile units and alleviates the computing load of fixed agents. However, confronting the intricate and diverse nature of computational tasks, strict adherence to the First-In-First-Out (FIFO) principle may lead to delays for preceding tasks, potentially resulting in incomplete processing or task discards. To address this, the paper proposes a prioritization strategy for a given task queue, dividing it based on service level agreements (SLA) to optimize critical performance within limited network capacities. Additionally, the computational offloading learning approach incorporates reinforcement learning methods from machine learning. This holistic approach aims to advance the efficiency and effectiveness of intelligent transportation systems in the face of evolving technological and societal demands.