Bi-Level Agent Social Trust(BLAST) Model: Simulating Systems of Trust from Internal and External Factors

In order to better model agent-agent interaction, a robust model for trust is necessary. Many computational trust models have been created in the past, but during the course of research, I’ve found none that model every facet of trust appropriately. In order to rectify this, I’ve devised a model based off of prior work in the field, and some ideas gleaned from sociology and neurobiology. The computational trust model presented in this paper takes into account both internal and external factors upon which trust is theorized to be based. This multi-directional approach covers agents that are naturally predisposed to trust, agents which trust based upon others’ role in a system, bond-making, experience, and the overall reputation of an agent in the system. It is my belief that combining all these factors of trust will create a model which is robust enough to approach the intricacies of human interaction.