Engineers and computer scientists have envisioned a world where cooperative mobile robot systems could be utilized in package delivery, agriculture, search and rescue, construction, remote exploration (deep sea, space, exoplanets, etc.), disaster relief, and a myriad of other applications. However, in reality, robots face constraints in their physical capabilities and must operate within boundaries set by cost, environment, and technology. My research is centered on establishing the theoretical and practical foundations for cooperative mobile agent systems, employing approaches such as theory, machine learning, blockchain, and other techniques. Specifically, I am interested in investigating the impact of communication, localization, sensing, mobility, and computation limitations on the ability of mobile agent teams to collaboratively solve problems. In this presentation, I will discuss work on two fundamental challenges in this field: message delivery and dispersed computing. I will cover my theoretical research on creating cooperative algorithms with optimal competitive ratios for various delivery problems, as well as our efforts to leverage machine learning and other techniques for task scheduling in dynamic environments. I will conclude the talk by addressing the remaining challenges and highlighting future research directions.