Web 2.0 applications are becoming more and more popular. In contrast with traditional applications, such as the search engine system and the traditional web portal, users can publish and access message/information instantly on the Web 2.0 platform, such as the Blogsphere, the Instant Message system, BBS etc. One property of the Web 2.0 that has attracted particular attentions is that of "community structure". Community is a sociologic term. The community is a group of people who form relationships over time by interacting regularly around shared experiences, which are of interest to all of them for varying individual reasons. The most important property of the community is the common interest and the united activities shared by the community members, i.e. a community is composed of the music fans of a swan.
The Internet is becoming tightly related to our daily life. Due to the popularity of the Internet services, especially the Web 2.0 applications, lots of social activities, such as traveling, shopping, chatting, and playing games, can be organized in the cyberspace. Since most of these activities are formed as dynamic communities, finding the hidden and meaningful community and charging its evolving progress has been the focus of many efforts. These researches can help build more user-friendly social software, and enlarge the "power of the crowds" for the Web 2.0 applications.
However, the Internet communities are self-organized based on diverse and complex ingredients. A logical community may be created by users for similar interests, intimate social relationships, geographical distance, past access behaviors of on the Web etc. At the same time, these rich ingredients related communities are created, evolving, superposing and vanishing all the time. Until now, there are not efficient methods to detect such rich ingredients related communities, what is more, to estimate the relationship among the diverse communities and to predict the propagation and evolution of the community on the Web.
The goal of this research is to analyze the fundamental phenomena and evolving characteristics of the community structure in the Web 2.0 systems. We will propose a kind of multi-dimensional model to define and identify the "meaningful" community on the Web 2.0. A new kind of complex adaptive multi-agent model will be used to measure the relationship between the communities. Some evolutionary computation methods will be explored to predict the evolving of communities on the Web2.0.