A Simulation Environment Predicting User Data Demand (UDD-Sim) in Upcoming Wireless Networks and Smart Cities

Inhalt:
In upcoming wireless networks, it is expected that automated orchestration and resource allocation will increase through Artificial Intelligence with various capabilities regarding both its action space and information input. Such multimodal models can process different data types and operate based on context. To train such systems, it is also necessary to test the system’s response to a broad spectrum of expected situations but also rare occurring events. However, access to suitable datasets is limited. This work presents a customizable framework named User Data Demand Simulation (UDD-Sim), which simulates internet traffic demand for mobile users and stationary devices for smart cities. The data demand simulation is initialized with heuristic analytical models based on available real data, however, the focus of this work is to share the structure and usability of the simulation environment with the scientific community, while intrinsic models are fully customizable. The user demand is split into several usage models allowing the consideration of different requirements like latency and throughput. Additionally, an exemplary simulation of a small village is presented, demonstrating the inherent visual inspection and evaluation capabilities of the introduced framework.