An Evolutionary Game Theory Model of Revision-Resistant Motivations and Strategic Reasoning

Abstract

Strong reciprocity and other forms of cooperation with non-kin in large groups and in one-time social interactions is difficult to explain with traditional economic or with simple evolutionary accounts. Reciprocity can be costly, while in many instances earning little or no benefit to the individual or its kin. In Ultimatum Games, for example. humans tend in one-shot anonymous interactions towards equal distributions of goods at high individual cost, often encouraged through retributive actions that result in significant personal cost. In this research, an agent-based genetic algorithm model is used to show that in a game similar to the Ultimatum Game, and of which an Ultimatum Game could be interpreted as a subgame, but where the past history of an agent's retributive actions is visible to other agents, strategies exhibiting strong reciprocity can evolve. This model is notable for its conservatism; it presupposes no special features in the structure of the population, relies solely upon potential benefits to kin and offspring, and requires only punishment (and not also reward) as an explanation of the behavior. The model also is consistent with a number of findings on the nature of emotions and related forms of motivation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2008
Accession Number
ADA493545

Entities

People

  • Craig Delancey

Organizations

  • State University of New York

Tags

Communities of Interest

  • C4I
  • Human Systems

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Cooperation
  • Delphi Method
  • Economics
  • Game Theory
  • Genetic Algorithms
  • Human Behavior
  • Military Research
  • Motivation
  • Motor Skills
  • New York
  • Psychology
  • Reasoning
  • Recreation
  • Simulations
  • Social Sciences

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Game Theory.
  • Strategic Security Studies

Technology Areas

  • AI & ML
  • Biotechnology