Developing Digital Twin of Plant Based on Emergent Intelligence Concept, Ontology and Multi-agent Technology

The article advances methodology of designing smart digital twins of plants (SDTP) as a multi-level complex adaptive system, integrating the capabilities of cyber-physical and ontology-customizable multi-agent systems for planning and simulation of plants’ growth stages, synchronized with stages of real crops. New multi-agent model and method of creating a SDTP are proposed, based on the use of the “emergent intelligence” concept which combine the advantages of centralized and self-organized resource management in a form of self-guided organization. The multi-level, multi-agent resource-service network model of the plant is developed, as well as a method of compensations, that ensures the quasi-optimality of plant plans. Ontologically configurable classes of agents and protocols of their interaction are presented. The new model and method are implemented in an ontology-customizable multi-agent system for resource scheduling and control of crops. Results of first experiments for wheat and broccoli are discussed and plans for future developments are given.