unit_cum_appx: Approximate Cumulative Unit Learning Curve Function

In learningCurve: An Implementation of Crawford's and Wright's Learning Curve Production Functions

Description

Provides the approximate cumulative time or cost required for units m through n (inclusive) using the Crawford unit model. Provides nearly the exact output as unit_cum_exact(), usually only off by 1-2 units but reduces computational time drastically if trying to calculate cumulative hours (costs) for over a million units.

Usage

unit_cum_appx(t, n, r, m = 1, na.rm = FALSE)

Arguments

t	time (or cost) required for the mth unit of production
n	The unit you wish to predict the cumulative time (or cost) to
r	learning curve rate
m	mth unit of production (default set to 1st production unit)
na.rm	Should NA values be removed?

Examples

library(learningCurve)
# An estimator believes that the first unit of a product will 
# require 100 labor hours. How many total hours will be required
# for 125 units given the organization has historically experienced
# an 85% learning curve?

unit_cum_appx(t = 100, n = 125, r = .85)
## [1] 5202.998

# Computational difference between unit_cum_exact() and unit_cum_appx() 
# for 1 million units

system.time(unit_cum_exact(t = 100, n = 1000000, r = .85))
##  user  system elapsed 
## 0.105   0.004   0.109

system.time(unit_cum_appx(t = 100, n = 1000000, r = .85))
## user  system elapsed 
##  0       0       0

learningCurve documentation built on May 2, 2019, 2:13 p.m.