Application-driven Resource Allocation for Resilient Multi-Agent Cloud Robotic Systems

Nowadays, robotic applications have been widely adopted to enhance the operational automation and performance of real-world cyber-physical systems used in Industry 4.0, agriculture, healthcare, and disaster management. Commonly, robotic applications are composed of latency-sensitive, data-heavy and compute-intensive tasks. Whereas, robots are constrained in their computational power and storage capacity. The thesis explores different approaches (evolutionary, game-theoretic and machine learning) to allocate computing resources from local robots, edge nodes and cloud servers for executing robotic applications. Consequently, it advances the state-of-the-art to develop automated systems with distributed resources by offering application-driven resource allocation mechanisms.