Performance, Multi-Application Scheduling with Intelligent Algorithms on Heterogeneous Resource Management Systems

Using software reconfigurable technology to provide multiple wireless access services for communication terminals synchronously has become a trend. It is required that UAVs have the ability to dynamically schedule waveform application under limited resource constraints. In this paper, we propose a heterogeneous resource runtime management framework and an intelligent algorithm, DRL-SDUS (Deep reinforcement learning and software-defined unmanned systems), which enables dynamic hardware and software resource scheduling for continuously arriving applications. Simulation results verify that the proposed algorithm can reduce the average latency time of multiple applications to improve the application deployment completion.