Toward the Development of a Predictive Computer Model of Decision Making During Uncertainty for Use in Simulations of U.S. Army Command and Control System
Abstract
In today's increasingly complex world of digital command and control, it is seldom obvious or intuitive how the introduction of new automation systems will affect the overall performance of battlefield command and control (C2) systems. Field observations can account for performance factors that are directly observable, such as rates of communication flow, rates of flow, and quality of incoming intelligence. However, what the human mind does under the influence of all these factors is not directly observable and is the subject of considerable experimentation. This research addresses this limitation through the development of predictive quantitative models of decision making during conditions of uncertainty such as exist in many aspects of human performance and certainly in battlespace management. Using Bayesian statistical approaches implemented through Partially Observable Markov Decision Processes (POMDP) that describe experiential decision processes moderated by Monte Carlo effects to account for performance variability, we are developing a series of computer simulations with the goal of predicting the quality of decisions possible from a given set of input conditions. These simulations are based on cognitive models being developed in a collaborative effort through a series of empirical studies that investigate human performance in a sequential decision making with uncertainty task using human subjects. Through this collaboration, the results of these studies are being applied at each stage of the research to predictive computer simulations of Army battlefield performance where battlefield automated C2 systems are involved. These simulations, when operational, will allow cognitive effects, such as predictive levels of effective decisions possible from a given set of circumstances, to be assessed as a battlefield metric.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2006
- Accession Number
- ADA443462
Entities
People
- Brian J. Stankiewicz
- Sam E. Middlebrooks
Organizations
- United States Army Research Laboratory