The Workshop on MAIN was held in conjunction with AAMAS 2013 on May 7, 2013 in Saint Paul, Minnesota, USA.
A network representation is very useful in both designing and analyzing multiagent systems. While simple theoretical models have been used for a while, like Erdos-Renyi random graphs, Watts-Strogatz small-world networks, and the Barabasi-Albert preferential attachment model, we are now seeing increasing application of detailed data-driven models, such as social contact networks obtained by combining multiple data sources, or networks induced from social media like Twitter and Facebook. Networks also offer a compact representation of complex (designed) multiagent systems, offering a new perspective on analysis of outcomes (e.g. simulation outcomes).
A network perspective also offers new opportunities for application of multiagent systems technology. Social science domains like epidemiology (including social epidemics like the spread of obesity, the spread of smoking, etc.) use social network data, but with very simple agent models, often consisting of a single variable. While it is well known that modeling human behavior is essential to these problems, little has been done on this front. Multiagent systems techniques have much to contribute here.
The goal of this workshop is to bring the network perspective to the forefront in the design and analysis of multiagent systems. We invite contributions from multiagent modelers, researchers in network science, and researchers in computational social science, with a focus on how network science might be used to improve the design of multiagent systems (by generating more realistic interaction structure, e.g.), and their analysis (some outcomes are best explained from a network perspective, e.g., time to consensus).
Contact: Samarth Swarup (firstname.lastname@example.org)