Analytics and typification of agents of a multi-agent model for a logistics system

Abstract

The article analyzes approaches to agent typification in multi-agent models in the context of combined transportation. The role-based, functional-architectural, behavioral approaches and the approach based on digital twins are considered. The necessity of hybrid classification taking into account business roles, functional purpose, level of autonomy, form of representation and level of decisions made is revealed. A typological matrix of agents is proposed, which makes it possible to formalize the composition of a multi-agent model at the design stage. The matrix can serve as a basis for the development of applied software solutions in logistics. The results of the study can be used in the creation of hybrid multi-agent systems combining physical and virtual components, which is especially relevant for the digital transformation of the logistics industry. The proposed approach can significantly increase the efficiency of logistics process management through decentralization and adaptability.