Game-theoretic Mechanisms for Group Decision Making
Abstract
While the Internet has revolutionized many aspects of human life, including commerce, adver- tising, social interactions, and education, it has not yet proved to be a force for the good in large scale decision-making. To address this problem, we start with the following premise: The mecha-nism designer for a societal decision making problem may not be able to enumerate the outcomes in the decision space or know their structure, and this decision space may be too big for most ordinal voting schemes. However, we assume that agents can still reason about their preferences and small groups of agents can negotiate over this space and collaboratively propose outcomes that appeal to many of them. Our goal is to design protocols that can aggregate societal preferences without a need to formally articulate the entire decision space and without every agent having to report ordinal rankings over this space. We will take two approaches, both rooted in game-theoretic ideas: 1. Scalable Deliberation via Small Group Interactions: The need for small groups is motivated by a practical consideration as well as a theoretical one. On the practical side, there is no onlineplatform, to the best of our knowledge, that has a successful history of large scale deliberation and decision making on complex issues; in fact, large online forums typically degenerate into vitriol and name calling when there is substantive disagreement among the participants. Thus, if we are to develop practical tools for decision making at scale, a sequence of small group deliberations appears to be the most plausible path. In recent work, we formalized some of the connections between sequential protocols for deliberation and axiomatic theories of bargaining for small groups. In this proposal, we seek to extend this preliminary work to develop an overarching formal theory for social choice via bargaining, with several real-world di?culties that we hadn t considered before. The ?rst is informational, i.e., what happensif participants only have Bayesian information about other participants? The second has to do with the total complexity of all the proposals that are generated during all the phases of a deliberation or bargaining process { can we model a synthesis step that allows similar proposals to be mapped and merged? We will also address the limit of how big a population sample we need for the outcome of a deliberative process on the sample to be representative of the full population. We will prototype and test our algorithms in a deliberation platformwe are building in collaboration with Jim Fishkin at Stanford.2. Markets and Mechanisms for Decision Making: Here, we assume that participants makea complex multi-dimensional decision (perhaps under budget constraints) by trading in an arti?cial market, with each individual getting an initial endowment of one unit of (fake) money. We have recently shown that the most natural market design in this setting results in highly sub-optimal outcomes, but more careful designs can lead to socially optimal market equilibria. Our current results are not e?ciently implementable, and this part of the proposed research will focus on developing practical implementations of our market, as well as approximationalgorithms for computing the equilibrium if the utility functions of each individual are known.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 15, 2019
- Source ID
- N000141912268
Entities
People
- Ashish Goel
Organizations
- Office of Naval Research
- Stanford University
- United States Navy