Problems of multi-agent systems and possible ways to solve them

This article discusses the history of the development of multi-agent systems, highlights their main advantages, and addresses the reasons why this technology has not become widespread: the complexity of the BDI agent model, FIPA standards, the lack of industrial of multi-agent systems development methodologies, etc. Various solutions are proposed to overcome the current situation, such as: creating a library of common of multi-agent systems design patterns, adapting existing approaches and technologies from the field of software engineering for the development of agent architectures and the overall multiagent system, and using large language models (LLaMa, Alpaca, etc.) for the implementation of intelligent software agents, etc.