The impact of diversity on performance of holonic multi-agent systems. (October 2016)

• Record Type: Journal Article
• Authors: Esmaeili, Ahmad; Mozayani, Nasser; Motlagh, Mohammad Reza Jahed; Matson, Eric T.
• Is Part Of: Engineering applications of artificial intelligence. Volume 55(2016:Jul.)
• Page Start: 186
• Page End: 201
• Publication Date: 2016-10
• Subjects: Multi-agent systems -- Holonic multi-agent systems -- Diversity -- Optimization; 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.2016.06.011
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: There are Numerous researches in the fields of social sciences and multi-agent systems that are dedicated to studying the role of diversity vs. individual capabilities in design and performance of human and artificial societies. This paper addresses more complex multi-level structures, namely holarchies, in multi-agent systems, and conducts several analyses on the effect of individuals' behavioral diversities in the constitution and performance of the whole system. The hypothesis is that diversity directly affects the structure and performance of a holonic organization. In order to check this, we first proposed a diversity-based holonic model for multi-agent systems and then carried out a set of simulations in which a holonic multi-agent system is employed in a function optimization application. In these simulations, we manipulated the behavioral diversity of the agents in various ways and carefully observed the changes that occur on the effectiveness of the system. These observations include the quality of solutions, the height and composition of the constructed systems, and the time efficiency of the model. The empirical results show that as the diversity of intra-holon agents increases, the height of holarchy is reduced while the sizes of the holons tend to increase. On the other hand, this increase in diversity results in an increase in the performance of the system in situations that diversity of the initial population is high. However, when the diversity of the initial population is low, the diversities inside the holons need to be adjusted around 60 percent in order to boost the performance.