Negotiation strategies for agents with ordinal preferences: Theoretical analysis and human study

Negotiation is a very common interaction between agents. Many common negotiation protocols work with cardinal utilities, even though ordinal preferences, which only rank the outcomes, are easier to elicit from humans. In this work, we focus on negotiation with ordinal preferences over a finite set of outcomes. We study an intuitive protocol for bilateral negotiations, where the two parties make offers alternately. We analyze the negotiation protocol under two settings: First, we consider the full information setting, where each party is fully aware of the other party's preference order. For this case, we provide elegant strategies that specify a sub-game perfect equilibrium. In addition, we show how the studied negotiation protocol almost completely implements a known bargaining rule. Second, we analyze the complementary no-information setting where neither party knows the preference order of the other party. For this case, we provide a Maxmin strategy and show that every pair of Maxmin strategies specifies a robust-optimization equilibrium. Finally, through a human study (N=150), we empirically study the practical relevance of our full information analysis to people engaging in negotiations with each other and/or with an automated agent using the studied protocol. Surprisingly, our results indicate that people tend to arrive at the equilibrium outcomes despite frequently departing from the proposed strategies. In addition, in contrast to commonly held beliefs, we find that an equilibrium-following agent performs very well with people.