Distribution String

DIST, (DIstribution STring) is an informal standard for an XML probability distribution container. It holds a randomized uniform partition of arbitrary size in a text string. With the appropriate add-in, it's particularly useful for modeling and simulation with Excel/VBA, allowing an entire distribution to be kept in a single cell.

As a text format, it can be treated as a constant--a consistent source to use as a model input, ensuring repeatability not just for different runs but on different systems, without any concern about compatibility of PRNGs.

Also, because DIST uses a uniform partition instead of a formula, a group of DISTs can preserve the connections between interrelated variables. This is vital for getting correct results from Monte Carlo Simulation.

DIST is described in Sam Savage's Flaw of Averages and supported by a variety of Excel and MS Project add-ins and other analytics products. DIST has a growing corporate user base, particularly financial and resource industry firms.

The DIST web site is at http://probabilityManagement.org

DIST uses a lossy compression scheme for the distribution vector and uses Base64 to convert the numbers to text for XML. The lossiness is not a concern if the distribution shape is fairly smooth; if it isn't it will be.

--MarcThibault 2010-10-05

CategoryFuzzy, CategoryVisualBasic, CategoryXml, CategoryStatistics