Our projects are being implemented in R at this time. R is world’s most widely used system for statistical computation and graphics. It is often the first choice of data scientists and supported by a vibrant and talented community of open source contributors. R is taught in universities and deployed in mission critical business applications. We find R to be an excellent platform for prototyping applications of a statistical nature and quickly delivering them to a wide audience in an open source environment.
Fault Tree Analysis
Since inception in the early 1960's fault trees have been used to map various cause and effect relationships across many fields of study. There have been many commercial implementations with some variation on the theme.
Here is a completely functional package for fault tree analysis on R.
As this project was approached colleagues requested that event trees also be implemented.
This project was started as an expository implementation of several functions supporting reliability analysis methods presented in "The New Weibull Handbook, Fifth Edition" by Dr. Robert B. Abernethy. After adopting the weibulltoolkit package much progress has been made toward a complete Weibull analysis application.
Stochastic Modelling for RAM Analysis
Reliability, Availabiilty and Maintainability analysis can be enhanced by discrete event simulation. This modelling handles time variation effects that cannot be discriminated with steady state analytical expressions such as used in fault tree modeling. This modelling can be essential for handling intermediate stored inventories used for backup operation during operating unit failures. It is also ideal for mapping the cases for varied-capacity unit operations through operations that include planned and unplanned downtime for individual units, while the overall system proceeds.
The stosim package is available on CRAN.
Jacob Ormerod presented this example analysis using the stosim package at the 8th IMA International Conference on Modelling in Industrial Maintenance and Reliability in July 2014 at Oxford University.
Multi-Echelon Supply Chain Optimization
This project has been inspired by Jorge Fukuda’s METRIC site. It is currently a simple expository implementation of example cases from several texts. Most noteworthy is “Optimal Inventory Modeling of Systems” by Craig C. Sherbrooke. This text describes the basic METRIC algorithm. Replacement parts are crucial to the maintainability of many complex systems. There is a constant challenge to balance the cost of holding spares against the benefit of avoiding waiting time on critical systems. The METRIC family of algorithms has been a key technology for this important supply chain consideration.
Here is the original development activity on the xmetric package for R. The project has been dormant for several years. Collaborators are welcome. Perhaps some folks from Aernnova for instance.