Multi-agent Approach to Community Detection in Complex Networks
نویسنده
چکیده
A multi-agent approach to community detection is studied. There are three objectives in this thesis. The first is to investigate how the parameters of the model affect the community structure. To investigate this, the parameters are swept one at the time and the results are then compared to each other. The second objective is to study how the initial values of the agents affect the community structure. This is studied by letting all parameters be fixed and vary the initial values. The third objective is to study how robust the model is to networks with negative links and networks with missing links. This is studied by fixing all parameters and replacing some of the positive links with negative links and comparing the outcome with the original communities. Removal of some links is then done in a similar way and compared to the original network. The study of the multi-agent approach has led to conclusions being drawn. The parameters ρ and α are sensitive and setting the initial values in a good way gives increased convergence speed. The final conclusions are that opinion dynamics with decaying confidence is a suitable model to networks that contains negative links while the robustness to missing links depends on the accuracy demanded by the application. Sammanfattning Multi-agent system används för att detektera kluster i komplexa nätverk. Det finns tre mål med uppsatsen. Det första är att undersöka hur parametrarna i den matematiska modellen p̊averkar klusterdetekteringen. Detta undersöks genom att parametrarna i modellen varieras och sedan jämförs resultaten. Det andra m̊alet är att studera hur begynnelsevärdena för agenterna p̊averkar klusterdetekteringen. Detta görs genom att fixera alla parametrar och variera begynnelsevärdena. Det tredje målet är att redogöra huruvida modellen är kompatibel med nätverk som inneh̊aller negativa och studera robustheten mot saknade länkar. Detta undersöks genom att fixera alla parametrar och sedan ersätta n̊agra positiva länkar med negativa länkar och sedan jämföra resultatet. Problemet med saknade länkar undersöks p̊a liknande vis. Denna studie har lett till följande slutsatser. Parametrarna ρ och α är känsliga. Genom att sätta begynnelsevärdena p̊a ett bra sätt kan man öka konvergenshastigheten. Den sista slutsatsen är att modellen är kompatibel med negativa länkar och beroende p̊a tillämpning s̊a är modellen även robust mot saknade länkar i nätverket.
منابع مشابه
A Multiagent Reinforcement Learning algorithm to solve the Community Detection Problem
Community detection is a challenging optimization problem that consists of searching for communities that belong to a network under the assumption that the nodes of the same community share properties that enable the detection of new characteristics or functional relationships in the network. Although there are many algorithms developed for community detection, most of them are unsuitable when ...
متن کاملOn the use of multi-agent systems for the monitoring of industrial systems
The objective of the current paper is to present an intelligent system for complex process monitoring, based on artificial intelligence technologies. This system aims to realize with success all the complex process monitoring tasks that are: detection, diagnosis, identification and reconfiguration. For this purpose, the development of a multi-agent system that combines multiple intelligences su...
متن کاملFault Detection and Isolation of Multi-Agent Systems via Complex Laplacian
This paper studies the problem of fault detection and isolation (FDI) for multi-agent systems (MAS) via complex Laplacian subject to actuator faults. A planar formation of point agents in the plane using simple and linear interaction rules related to complex Laplacian is achieved. The communication network is a directed, and yet connected graph with a fixed topology. The loss of symmetry in the...
متن کاملA Fast Approach to the Detection of All-Purpose Hubs in Complex Networks with Chemical Applications
A novel algorithm for the fast detection of hubs in chemical networks is presented. The algorithm identifies a set of nodes in the network as most significant, aimed to be the most effective points of distribution for fast, widespread coverage throughout the system. We show that our hubs have in general greater closeness centrality and betweenness centrality than vertices with maximal degree, w...
متن کاملCommunity Detection using a New Node Scoring and Synchronous Label Updating of Boundary Nodes in Social Networks
Community structure is vital to discover the important structures and potential property of complex networks. In recent years, the increasing quality of local community detection approaches has become a hot spot in the study of complex network due to the advantages of linear time complexity and applicable for large-scale networks. However, there are many shortcomings in these methods such as in...
متن کاملIdentifying overlapping communities using multi-agent collective intelligence
The proposed algorithm in this research is based on the multi-agent particle swarm optimization as a collective intelligence due to the connection between several simple components which enables them to regulate their behavior and relationships with the rest of the group according to certain rules. As a result, self-organizing in collective activities can be seen. Community structure is crucial...
متن کامل