Expanded Fermi Solution for Risk Assessment

seed value

0

seed repeatable random numbers

data

generate

clear

random trials to generate

500

minima

maxima

n

500

800

n multiplier

1

p

1

0.2

0.7

p

2

0.2

0.6

p

3

0.2

0.5

p

4

0.6

0.8

p

5

0.8

0.9

p

6

0.9

1.

In a "standard Fermi solution", one multiplies guesses of several factors to reach an estimate. This Demonstration uses an expanded version of the method, originally developed for microbial risk assessment, in which the ranges of the factors are entered instead of their values. Guesses are then generated repeatedly using random factor values within the entered ranges and the resulting estimates are plotted as a histogram on which their distribution's curve, which is approximately lognormal, is superimposed. The mode of this distribution is considered the "best estimate". It is also calculated analytically and both values are displayed on the plot.