Analysis of System Behavior through Cognitive Architectures

Abstract

This paper focuses on utilization of multiagent systems and cognitive architectures for analysis of financial market dynamics. Sloman's H-Cogaff cognitive architecture is utilized as a blue-print to design trader architecture (Sloman 2002). The trader architecture combines a Markov-based bias mechanism and an XCS classifier based learning mechanism. An agent-based artificial stock market is designed to observe trader behavior and market price dynamics. Initial results reveal that the bias mechanism affects price dynamics and results in deviation of the stock price from the efficient market price.

Department(s)

Engineering Management and Systems Engineering

Keywords and Phrases

Cognitive Architectures; Agent-Based Simulation; Artificial Stock Market; Bias Mechanisms; Learning Classifier Systems

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Publication Date

01 Jan 2007

This document is currently not available here.

Share

 
COinS