Date of Award

12-2019

Degree Type

Thesis

Degree Name

Master of Science in Analytics and Modeling

Program

Analytics and Modeling

First Advisor

Rick Gillman

Abstract

In game theory, buyer-seller games rarely utilize a negotiating third party. Any negotiations are typically conducted by the buyer and seller. This study, motivated by the real estate market, uses sequentially and simultaneously played game models to explore the influence a self-interested, negotiating, third party has on player payoffs. For the sequential model, a game tree is utilized to demonstrate player actions, preferences, and outcomes. The weak sequential equilibrium is calculated using Gambit[1] and shows optimality in player payoffs to exist when the seller’s and realtor’s strategies align according to the current market. For the simultaneous model, expected payoff functions for each of the three players are constructed. PlatEMO[2], a MATLAB extension, is used to simultaneously maximize the players’ functions using multi-objective optimization evolutionary algorithms. The Pareto-optimal front is found, consisting of all non-dominated solutions in the objective space. Similar to the sequential model, optimal outcomes exist when seller and realtor strategies align. Findings from both models suggest a self-interested negotiating third party is largely unnecessary and only has negative impact on player payoffs.

Download

Share

COinS