Computing a Model for Asset Management with Risk - Global Drinking Water Management and Conservation - page 182

Fig. 8.14 A nonlinear,
logistic S function
approximating S(x) in City B,
over the range 25 - 40

Fig. 8.15 A nonlinear,
logistic S function
approximating S(x) in City B,
over the range 15 50

-

investment cost in this city is $152,169, which is determined by taking half the sum
total of failure costs during the sample period.

8.4.4.1 Determination of S(x)

The cost of failure in City C was regressed on all available variables. However, only
size and depth variables were signi cant. After correcting multicollinearity, heter-

oscedasticity, and non-normality, a Robust LAE regression was then performed
utilizing all available variables and this revealed that all the variables, with the