A reputation-based model for semi-competitive multi-agent systems

Abstract

In an e-business application, an agent chooses its partners according to their reputations while maximising its utility. In this paper, we propose a mathematical model to compute the reputation of an agent. We improve existing models by including additional sources of information in reputation computing. Furthermore, each source of information has a weight that reflects its importance. So we also propose a new approach for computing the weight of each source of information. We show that our model is valid since while adding new sources of information, agents behave as expected.