Cog-Trust

From Trustwiki

Jump to: navigation, search


The Cog-Trust project


Cog-Trust is an open source project for an agent-based Java framework designed on the Computational model of Socio Cognitive Theory of Trust.


Details

Cog-Trust realizes an hybrid agent architecture: a deliberative module is coupled to classifiers and a set of fuzzy based decision-making modules used for trust assesment.

Cog-Trust provides a multi-agent simulator to test different strategies of delegation based on trust playing a repeated trust game. Trustors compete in succesfully delegating the assigned task to agent choosen within a dynamic population.

Cog-Trust also provides features for designing environments, populations and built-in infrastructues for agent communication, contract-net and events notifications.

User can add their own strategy of delegation to the predefined set of trustor and observe how it performs in the Cog-Trust simulation. Tools for data analysis and visual trust network monitoring are also available.

Features:

  • BDI agent implementation using Jason.
  • CArtAgo environment modeling and programming.
  • JFCM: Fuzzy Cognitive Maps Java framework.
  • Data Logging.
  • Matlab scripts for data analysis.
  • Visual monitoring of trust networks.

Developers:

Scientific supervisor:

Screenshots

View of the trust network


Bar scores


Chart scores

Documentation

Cog-Trust actually runs as a Jason application running inside a CArtAgO environment. All you need to get started with the Cog-Trust framework is:

  • Download and extract the leatest ZIP release of Cog-Trust from here.
  • Download and extract the leatest release of the Jason platform from here.
  • Run the Jason platform, then load and run the example file CogTrust.mas2j from the CogTrust directory.


If you are interested in working on Cog-Trust, the following resources are available:


For further support you can email matteo.venanzi(at)istc.cnr.it

This page was last modified on 2 November 2010, at 23:24. This page has been accessed 2,382 times.