The current model for LLM services involves receiving the service through cloud sources for all users. However, this solution is expected to have scalability problems. In this project, we intend to test the performance of the alternatives for offering LLM Services over Edge Computing.
Air computing is offering edge computing services over aerial vehicles such as UAVs.
We have developed a new simulator for Air Computing: AirCompSim.
This project involves adding energy consumption and constraints features to AirCompSim and testing the new simulator with energy aware resource allocation/trajectory computation for air computing.
Sudden natural disasters such as wildfires, earthquakes, hurricanes require efficient management techniques to reduce the losses. Some of these disasters trigger new ones that may deepen the damage or widen the locations prone to the disaster. For example, a major earthquake may trigger a number of earthquakes which may lead to more disaster locations. A wildfire may change its direction of spread. In this project, we will call this type of disaster “an invasive disaster”, i.e. the disaster locations and the loss rate at each disaster location may change as time progresses.