AI-agent-based modeling for Supervision a System of Systems for Mushroom Harvesting

Mushroom harvesting is considered a complex task as it requires the cooperation of different available resources, including the cooperation between human operators and harvesting assistance robots, in order to obtain an efficient process. In this context, we present in this paper an agent-based modeling (ABM) to supervise the system of systems (SoS) of the harvesting process, considering its static and dynamic representations, in case of changes in goals and mission allocation. Furthermore, we propose the modeling of the agents with an AI approach based on reinforcement learning (RL), a simulation of the mushroom harvesting process has been finally implemented to test the RL approach for each elementary component system (CS) at the low level of the SoS.