R: Optimal Experimental Designs for Accelerated Life Testing

ALTopt-package {ALTopt}

R Documentation

Optimal Experimental Designs for Accelerated Life Testing

Description

Creates the optimal (D, U and I) designs for the accelerated life testing with right censoring or interval censoring. It uses generalized linear model (GLM) approach to derive the asymptotic variance-covariance matrix of regression coefficients. The failure time distribution is assumed to follow Weibull distribution with a known shape parameter and log-linear link functions are used to model the relationship between failure time parameters and stress variables. The acceleration model may have multiple stress factors, although most ALTs involve only two or less stress factors. ALTopt package also provides several plotting functions including contour plot, Fraction of Use Space (FUS) plot and Variance Dispersion graphs of Use Space (VDUS) plot. For more details, see Seo and Pan (2015) <doi:10.32614/RJ-2015-029>.

Details

Package:	ALTopt
Version:	0.1.2
Authors@R:	as.person(c( "Kangwon Seo <seoka@missouri.edu> [aut, cre]", "Rong Pan <rong.pan@asu.edu> [aut]" ))
Depends:	R (>= 3.0.0)
License:	GPL-3
LazyData:	true
Imports:	cubature (>= 1.0), lattice (>= 0.20), methods
Built:	R 3.6.1; ; 2019-12-12 12:30:00 UTC; windows

Index:

alteval.ic              Design evaluation with interval censoring.
alteval.rc              Design evaluation with right censoring.
altopt.ic               Optimal design with interval censoring.
altopt.rc               Optimal design with right censoring.
compare.fus             Comparing designs using FUS
compare.vdus            Comparing designs using VDUS
convert.stress.level    Coding and decoding stress level
design.plot             Design plot.
pv.contour.ic           Contour plot of prediction variance for a
                        design with interval censoring.
pv.contour.rc           Contour plot of prediction variance for a
                        design with right censoring.
pv.fus.ic               FUS (Fraction of Use Space) plot for interval
                        censoring.
pv.fus.rc               FUS (Fraction of Use Space) plot for right
                        censoring.
pv.vdus.ic              VDUS (Variance Dispersion of Use Space) plot
                        for interval censoring.
pv.vdus.rc              VDUS (Variance Dispersion of Use Space) plot
                        for right censoring.

Author(s)

Kangwon Seo [aut, cre], Rong Pan [aut]

Maintainer: Kangwon Seo <seoka@missouri.edu>

References

Seo, K. and Pan, R. (2015) ALTopt: An R Package for Optimal Experimental Design of Accelerated Life Testing. The R Journal, 7(2), 177-188.

Monroe, E. M., Pan, R., Anderson-Cook, C. M., Montgomery, D. C. and Borror C. M. (2011) A Generalized Linear Model Approach to Designing Accelerated Life Test Experiments, Quality and Reliability Engineering International 27(4), 595–607

Yang, T., Pan, R. (2013) A Novel Approach to Optimal Accelerated Life Test Planning With Interval Censoring, Reliability, IEEE Transactions on 62(2), 527–536

Examples


# D optimal design of two stress factors with right censoring.
Design.D <- altopt.rc("D", 100, 100, 2, 1, formula = ~ x1 + x2 + x1:x2,
coef = c(0, -4.086, -1.476, 0.01))

design.plot(Design.D$opt.design.rounded, x1, x2)

pv.contour.rc(Design.D$opt.design.rounded, x1, x2, 100, 2, 1,
formula = ~ x1 + x2 + x1:x2, coef = c(0, -4.086, -1.476, 0.01),
useCond = c(1.758, 3.159))

FUS.D <- pv.fus.rc(Design.D$opt.design.rounded, 100, 2, 1,
formula = ~ x1 + x2 + x1:x2, coef = c(0, -4.086, -1.476, 0.01),
useLower = c(1.458, 2.859), useUpper = c(2.058, 3.459))

# U optimal design of two stress factors with right censoring.
Design.U <- altopt.rc("U", 100, 100, 2, 1, formula = ~ x1 + x2 + x1:x2,
coef = c(0, -4.086, -1.476, 0.01), useCond = c(1.758, 3.159))

design.plot(Design.U$opt.design.rounded, x1, x2)

pv.contour.rc(Design.U$opt.design.rounded, x1, x2, 100, 2, 1,
formula = ~ x1 + x2 + x1:x2, coef = c(0, -4.086, -1.476, 0.01),
useCond = c(1.758, 3.159))

FUS.U <- pv.fus.rc(Design.U$opt.design.rounded, 100, 2, 1,
formula = ~ x1 + x2 + x1:x2, coef = c(0, -4.086, -1.476, 0.01),
useLower = c(1.458, 2.859), useUpper = c(2.058, 3.459))

# Comparing D and U optimal design.
compare.fus(FUS.D, FUS.U)