Misrepresentation-proof strategies in poly-agent systems

Abstract False representation prevention is an important issue in the management of poly-agent systems when a system manager needs to collect information from agents. This paper aims at proposing strategies to prevent false representation effectively and efficiently. We suppose that the manager can know the truth by inspecting but can not inspect all reports due to the high cost of inspection, and the manager will announce a penalty rule and a probability of doing inspection for preventing false representation. After formulating three penalty rules, we obtain the minimal inspections required for inducing true information under each penalty rule.