Oranits: Mission Assignment and Task Offloading in Open RAN-based ITS using Metaheuristic and Deep Reinforcement Learning
Researchers have introduced Oranits, a new system designed to optimize mission assignment and task offloading in Open Radio Access Network (Open RAN)-based intelligent transportation systems (ITS). The system addresses limitations in existing studies by explicitly considering mission dependencies and the costs associated with offloading tasks to edge servers. Oranits employs a two-pronged optimization strategy: a metaheuristic-based Chaotic Gaussian-based Global ARO (CGG-ARO) algorithm and a Multi-agent Double Deep Q-Network (MA-DDQN) deep reinforcement learning framework. Simulations indicate that MA-DDQN significantly outperforms CGG-ARO and baseline methods, improving mission completion rates by 11.0% and overall benefit by 12.5%. AI
IMPACT Enhances efficiency and adaptability of AI task processing in intelligent transportation systems.