Showing papers on "Uncertainty analysis published in 1972"

First-order uncertainty analysis of soil deformation and stability

Abstract: THIS PAPER REPORTS ON PROBABILISTIC METHODS OF ANALYSING THE DEFORMATION AND STABILITY OF SOIL FOUNDATIONS OR EMBANKMENTS. THE SPECIFIC PROCEDURES DESCRIBED WERE DEVELOPED ORIGINALLY FOR AN EARTH DAM IN RESPONSE TO QUESTIONS AS TO WHETHER ADDITIONAL SOILS EXPLORATION SHOULD BE UNDERTAKEN IN THE LIGHT OF THE OBSERVED DISPERSION IN THE DATA ALREADY AVAILABLE. METHODS OF ESTIMATING THE TOTAL UNCERTAINTY FACED BY THE ENGINEER PREDICTING THE DEFORMATIONS AND STABILITY OF SOIL STURCTURES OR FOUNDATIONS ARE DISCUSSED. THIS TOTAL UNCERTAINTY INCLUDES UNCERTAINTY IN THE SOILS PROPERTIES (CAUSED BY LIMITED SPATIAL SAMPLING AND BY TESTING VARIABILITY), SUBSEQUENT UNCERTAINTY IN THE PREDICTED QUANTITIES (E.G., DEFORMATIONS), AND UNCERTAINTY IN THE ENGINEERING THEORY OR ALGORITHM USED TO PREDICT THESE QUANTITIES. THIS PAPER DISCUSSES A PRACTICAL FORMULATION OF SUCH PROBLEMS WHICH PRODUCES A FIRST - ORDER UNCERTAINTY ANALYSIS. SECONDLY, IT PRESENTS A NUMERICAL EXAMPLE WHICH INVOLVES EMBANKMENT STABILITY ANALYSIS AND WHICH MAKES USE OF REGRESSION ANALYSIS TO ESTABLISH SPATIAL SOILS PROPERTY VARIATIONS AND THE UNCERTAINTY IN THEIR ESTIMATES. /TRRL/

Confidence in Estimated Airframe Costs: Uncertainty Assessment in Aggregate Predictions,

Abstract: : The study develops and applies a method for quantitative analysis of the uncertainty inherent in certain predictive models involving a number of regression equations Specifically, it analyzes the uncertainty associated with the Rand airframe cost estimating model

Response uncertainty and autonomic-behavioral integration

Abstract: During the past several years, I have been engaged in a program of psychophysiological research designed to explore the integration of autonomic and somatic-behavioral responses. These investigations have clearly shown that certain elementary manipulations of simple overt behaviors consistently produce autonomic-activational correlates in human subjects. However, these concomitant autonomic events do not occur on all occasions of overt behavior, and one difficulty we have experienced is, quite literally, “coming to terms’’ with these phenomena. The differential participation of autonomic responses in somatic-behavioral events is clearly demonstrated by the psychophysiological effect of “activational peaking.” The effect has been produced in a wide variety of animal conditioning and human learning situations, employing different measures of activation, e.g. galvanic skin response (GSR), heart rate (HR), and electroencephalographic (EEG) desynchronization. Briefly, the “activational peaking” effect consists of the following sequence of events. During the initial stages of behavioral conditioning, activational responses to the conditioned stimulus (CS) show systematic, sometimes trial-by-trial, increments, until some stable level of conditioned response (CR) performance has been achieved. Subsequent activational responses to the CS decrease with further conditioning trials, and the well-established behavioral CR may occur without any measurable autonomic antecedents. It should be noted that some conditioning studies have shown similar “peaking” effects in the C R measure, itself, and it would appear that such “sensitization” occurs when the CR is affected by changes in general activation. However, the general point of interest here is that the performance of initial CRs or early competing responses to the CS are preceded by nonspecific activational changes, whereas the performance of the well-established CR may occur without apparent autonomic involvement. The primary concern of this research program, then, has been the identification of the behavioral conditions that determine the extent of autonomic participation in somatic-behavioral events. The results have shown the following relationships. First, whenever a simple overt behavior (e.g. a finger movement) is associated with a stimulus, the extent of autonomic or activational involvement is greatly increased. Signal stimuli (i.e., those associated with behavioral responses) frequently produce multiphasic activity and tonic shifts in activation level, in anticipation ofthe overt behavioi. However, activational responses to both signal and nonsignal stimuli habituate over trials.* Second, activational responses habituated to signal stimuli reemerge when subjects are required to change the overt behavior associated with these same stimuli along some simple quantifiable dimension. The degree of reemergence of the habituated activational responses is a direct function of instructed behavioral ~ h a n g e . ~ Third, the level of activational responses to signal stimuli across habituation trials is a direct function of the number ofpossible behavioral responses permitted by instructions to human subjects. For example, subjects instructed to say any number from one to ten after each signal stimulus demonstrate significantly larger GSRs to the stimuli than subjects instructed to say any number from one to five. These last subjects, however,