Application of a Genetic Algorithm and Multi Agent System to Explore Emergent Patterns of Social Rationality and a Distress-Based Model for Deceit in the Workplace
Abstract
Organizations rely on the honest operation of its members, but in an environment where individual members cannot be observed the opportunity for individuals to lie can lead to dishonest choices (Grover, 1993). This thesis created and applied a computer-based Genetic Algorithm and Multi Agent System in order to test the predictions of Dr. Steven Grover's distress-based model of the antecedents of lying in organizations. Grover's model blends self-interest theories and uses role theory to identify potential antecedents to lying. The created system provided agents that encountered situations of distress such as those described by Grover's model. The agents actions were then observed and compared to Grover's hypothesis that an individual's skill will be inversely proportional to his frequency of lying. Social rationality has been shown to emerge in simple self-interested agents. A hypothesis that in an environment where an organization and its members are independently self-interested, the frequency of organization members lying will be inversely proportional to the magnitude of feedback provided to the organization was tested. The results support both Grover's hypothesis and the hypothesis on social rationality. Self-interest individuals with higher skills lied less than individuals with lower skills. Also, self-interested individuals lied less in the presence of a higher magnitude of negative organizational feedback.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2008
- Accession Number
- ADA483642
Entities
People
- Jacob F. Davis
Organizations
- Naval Postgraduate School