Agent-based scheduling refers to applying Multi-Agent System (MAS) to model the scheduling environment, where intelligent agents cooperate to determine the task scheduling, so as to achieve scheduling objectives autonomously. To date, agent-based scheduling approaches have been developed for multiple domains, such as cloud computing, manufacturing systems, and smart grids. However, with the applications of the Internet of Things (IoT) and advanced communication technologies, real-world scheduling approaches are challenging to achieve satisfactory performances due to four challenges, which are (1) the expansion of the scheduling scale, (2) the involvement of uncertainties during the scheduling, (3) multiple real-world constraints, and (4) system resilience and robustness.