STSM WG5 (Eva Zupancic)

Start Date: 2012-04-15

End Date: 2012-05-15

Host Institution: University of Tallinn

Host Country: Estonia

Home Institution: University of Ljubljana

Home Country: Slovenia

Description:

It is still an open question how can a group of agents/services best organize themselves and interact to solve a given problem. Complex problems frequently can be solved by rearranging a group of agents. When rearranging the organization, trust between agents is an important factor that should be taken into account.

In the project, we will first try to analyze and find out what kinds of virtual organizations are suitable for pervasive computing, with the emphasis on eHealth systems. We will perform the analysis using the approach of agent-oriented modeling co-authored by the hosting professor, which supports analyzing and modeling a given problem domain from three balanced and interrelated viewpoint aspects,  interaction, information, and behavior  and from three horizontal abstraction layers: analysis, design, and platform-specific design. Based on the resulting structures for eHealth virtual organizations, we will then develop a prototypical multi-agent system. The interactions in pervasive intelligent environments also include exchange of data, acquired/measured by e.g. wireless sensors. To overcome hardware constraints, middleware solutions should be used to implement intelligent agents. According to agent-oriented modeling, such middleware solutions will be modeled and designed as services.

Further, we will also analyze the properties and organizational structures of current intelligent agent-based systems that are used to support healthcare processes. We will suggest the improvements, such that the agents would be able to solve the problem at hand in a more efficient and accurate way. The structure of a prototypical multi-agent system to be worked out will include the trust model, derived from Qualitative Assessment Dynamic. The integrated trust management system will provide the agents with an incentive mechanism and will consequently increase the quality of problem solutions.

Report

SetPageWidth