Measuring Errors in Operational Analysis Assumptions

Operational analysis, an area of study first defined in the computer science field, has been used in the analysis of systems performance. System performance measures for a specific set of output data are obtained using operational analysis formulas derived from assumptions which are verifiable by the observed data. This paper gives relationships which may be used to quantify the errors in these assumptions. Additionally, basic propositions are given which help in understanding operational analysis assumptions. These propositions are used in developing correction terms which can be used to adjust performance measures so that their values are exact for a set of data no matter how much the assumptions used in deriving the performance measure relations are violated.