CONTROL OF A DISCRETE STOCHASTIC PROCESS AS A FUNCTION OF THE COSTS FOR MAKING CORRECTIVE ACTIONS.
Abstract
This research examines man's ability to control a discrete stochastic process. The cost (C1) for correcting the process before it exceeded an arbitrary tolerance limit served as the experimencal variable. The cost (C2) for correcting this process after it exceeded the tolerance limit was fixed. For low C1 costs the human controller achieved optimum control (i.e., minimized control cost) within four 100-cycle trials. For high C1 costs the human controller achieved a level of control equivalent to that of the optimum controller on the first 100-cycle trial. An additional requirement to serve as a statistical sensor was imposed in controlling the process under low C1 cost conditions. The human controller's response to this requirement was appropriate. The implications of these results to the design of discrete stochastic process controllers is discussed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1963
- Accession Number
- AD0432826
Entities
People
- John P. Hornseth
- Walter J. Huebner
- William H. Pearson
Organizations
- Air Force Research Laboratory