Multi-Attribute Decision Making in a Bidding Game with Imperfect Information and Uncertainty

Abstract

Human preferences or attitudes towards risk should play a vital role in a decision making task with imperfect information and uncertain outcomes. We introduce a method to characterize human preferences and how they are integrated into the decision making process of a complex probability-based card bidding game. When assessing the preferences, a utility-to-preference (UP) function is devised for easier mapping between preferences and how much a player is willing to bid. Using the developed approach, we can better identify how different human preferences and their interaction affect the game outcomes. We focus on a highly addictive poker game that has become a multi-billion dollar internet business. The method was evaluated through data obtained from two decision makers (DMs) with different expertise. The integrated decision making process was designed and automated through Monte Carlo simulation. The results show that different preferences to the multi-attributes can lead to different profit outcomes. The results can further serve as a basis to identify vulnerable populations1 for the socio-technical online bidding game.