Individual Differences and Group Aggregation for Complex Decision Making Tasks
Author | : Sheng Kung Michael Yi |
Publisher | : |
Total Pages | : 97 |
Release | : 2011 |
ISBN-10 | : 1124511148 |
ISBN-13 | : 9781124511146 |
Rating | : 4/5 (146 Downloads) |
Download or read book Individual Differences and Group Aggregation for Complex Decision Making Tasks written by Sheng Kung Michael Yi and published by . This book was released on 2011 with total page 97 pages. Available in PDF, EPUB and Kindle. Book excerpt: The phenomenon of the `wisdom of the crowds' refers to the finding that the aggregate of a set of proposed solutions from a group of individuals performs better than the majority of individual solutions. Most work thus far in this field has focused on demonstrating this effect in simple environments where individual differences can easily be seen to contribute to a group consensus that lies close to the truth. In this research, we look at the problem of describing individual differences and performing group aggregation in two different types of complex decision-making tasks. Chapters 1 and 2 of this dissertation looks at application of wisdom of the crowds to combinatorial optimization-style problems, including the minimum spanning tree problem (MSTP) and traveling salesman problem (TSP). The first chapter looks at the wisdom of the crowds in a traditional fashion aggregating between individuals, while the second chapter investigates the use of aggregation techniques in a within-individuals scenario. Through both chapters, we describe a method for aggregating multiple problem solutions into single proposal solutions whose performances are among or exceed that of the solutions that generated them. Chapters 3 and 4 look at two different variants of the bandit problem, a classic sequential decision-making problem. The former of these focuses on the restless variant, while the latter focuses on a multiple-allocation variant. Two different approaches to describing behavior in these two branches of the bandit problem are provided, with additional brief exploratory discussion of aggregation analyses using the methods developed.