The Information Flow Problem in multi-agent systems. (April 2018)

• Record Type: Journal Article
• Authors: Búrdalo, Luis; Terrasa, Andrés; Julián, Vicente; García-Fornes, Ana
• Is Part Of: Engineering applications of artificial intelligence. Volume 70(2017:Oct.)
• Page Start: 130
• Page End: 141
• Publication Date: 2018-04
• Subjects: Multi-agent systems -- Communication strategies -- Agent communication -- Information exchange; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2018.01.011
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: One of the problems related to the multi-agent systems area is the adequate exchange of information within the system. This problem is not only related to the availability of highly efficient and sophisticated message-passing mechanisms, which are in fact provided with by current multi-agent platforms, but also to the election of an appropriate communication strategy, which may also greatly influence the ability of the system to cope with the exchange of large amounts of data. Ideally, the communication strategy should be compatible with how the information flows in the system, that is, how agents share their knowledge with each other in order to fulfill the system-level goals. In this way, MAS designers must deal with the problem of analyzing the multi-agent system with respect the communication strategy that best suits the way the information flows in that particular system. This paper presents a formalization of this problem, which has been coined as the Information Flow Problem, and also presents a complete case study with an empirical evaluation involving four well-known communication strategies and eight typical multi-agent systems.