COMING SOON! PQDT Open is getting a new home!

ProQuest Open Access Dissertations & Theses will remain freely available as part of a new and enhanced search experience at www.proquest.com.

Questions? Please refer to this FAQ.

Dissertation/Thesis Abstract

Exploring Social Discrimination with Artificial Life
by Reedy, Cara, Ph.D., Rensselaer Polytechnic Institute, 2020, 78; 28256168
Abstract (Summary)

The aim of this dissertation research was to begin approaching a larger goal of studying social discrimination showing the experimental potential of artificial life simulation in exploring problems involving both genetics and social behavior. Artificial life simulations can be used to conduct experiments that cannot be carried out on real humans, are easily manipulated and repeated, can collect detailed genetic data, and have an accelerated time-scale that can be used to study effects over many generations.

The first stage of this research was the creation and initial testing of an artificial life package capable of multiple kinds of environments, Garden, and a custom genetics package, DGEANN, which supports multiple genetic models. Experimentation showed that diploid genetics worked well in Garden and were used for the following experiments. In the second stage, pilot studies began introducing social behavior involving family relationships and inbreeding, focusing on how mating behavior and inbreeding avoidance influenced the population's genetic diversity, behavior, and long-term survival.

In the last stage of the research presented here, mating choice and preference for similar or dissimilar mates was examined as a simple form of discrimination. Mate similarity preferences evolved in environments that included genetic disorders and migration to another population, which influenced the direction in which those preferences evolved. The higher migration level led to agents preferring more similar mates, as did the higher initial percentage of genetic disorder carriers. The artificial life simulation shown in this work can be applied to other real-world problems involving both social and genetic components.

Indexing (document details)
Advisor: Hendler, James
Commitee: Si, Mei, Sun, Ron, Spector, Lee
School: Rensselaer Polytechnic Institute
Department: Cognitive Science
School Location: United States -- New York
Source: DAI-B 82/8(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computer science, Biology, Artificial intelligence
Keywords: Agent-based simulation, Artificial life, Evolving neural networks, Social simulation
Publication Number: 28256168
ISBN: 9798582525950
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest