Trust strategies for ART Testbed

Recently several trust models have been developed.

However, different trust models make different assumptions

about agents’ environments, making it difficult to evaluate and

compare different models. An accurate examination requires

different models to be tested on the same environment and

evaluated using precise metrics. For this reason, we adopt wellknown

trust concepts and formulate them as agent strategies

that can be applied in the Agent Reputation and Trust (ART)

Testbed Simulation Environment. We devise trust strategies that

can be applied at different stages of agents’ interactions. We study

different metrics for each stage and compare various strategies

using these metrics.