Posted by Theory of Complex Work
From the Journal of Cost Analysis and Parametrics: Volume 10 | Issue 3 | November 2022
Downloadable File: JCAPv10i3-TheoryofComplexWork-Larsen
Abstract:
The manufacturing learning curve graph depicts the direct labor hours per unit of production plotted in its manufacturing sequence. When plotted with log/log scales, learning curves tend to follow a linear downward trend as shown in figure 1.
The power law, y(x) = a xb, has been found to describe the learning curve’s trend, with y(x) = hours(unit), x = sequential unit number, a = the hours at unit one, and b the slope exponent. The power law is transformed to a linear form by taking its logarithm: ln(y) = ln(a) + b ln(x). To summarize a learning curve’s statistics, a least squares regression can be performed on the logarithms of learning curve data, (ln(y1), ln(1)); (ln(y2), ln(2)); . . . , (ln(yn), ln(n)). The regression estimate of a is termed the theoretical number one. By convention, the slope of a learning curve is 2b. The learning curve is sometimes transformed to a cumulative average.
Author:
Harry Larsen attended the University of Washington, graduating with a mathematics degree in 1966. After graduation, he was hired by the Boeing Company as a production estimator. He began building planning systems for Boeing operations on the company’s new supercomputer. Many systems and twenty years later he became the Business Planning Manager of Boeing Marine Systems, a division of the Boeing Company, reporting to the division’s general manager. He retired in 2002.