LLM-Based Multi-Agent Systems
Ravi Sharma, Dinesh Kumar Sahu, Rajnish Chaturvedi · Advances in computational intelligence and robotics book series · 2025
Large language model based multi agent system (LLM based MAS) is undoubtedly the most promising approaches towards artificial intelligence (AI). To build MAS, researchers utilize two major capabilities of LLMs viz. logical reasoning and effective planning. The existing literature related to MAS is indeed scattered and there is an urgent need to collect the information at one place to gain deeper insights. With this motivation this chapter was designed with the primary focus on providing a detailed overview of LLM based MAS. The chapter is divided into five major sections; construction of LLM-MAS, its key capabilities, its current applications with detailed examples, intrinsic limitations and challenges and finally a brief discussion about future trends in LLM based MAS research. Through this chapter we want our readers to gain a better understanding of this technological advancement which will enable us to unlock unseen opportunities in various public sectors.