A Method for the Statistical Evaluation of Small Subsystem Performance

This paper proposes a method for the statistical evaluation of small subsystems performance. The method requires that performance data from a small subsystem be analyzed using a multiple linear regression method; this has been programmed for a digital computer making use of a step-wise solution. The result along with predetermined component variation is then used in a computer programmed Monté Carlo simulation to develop both the output distribution and descriptive statistics. Experimental results from a small, one-core impulse-switched memory system are presented. Other applications include the analysis of high speed computer circuits; one application integrated circuitry, offers great promise.