The transportation modeling industry has several vexing problems: we are constantly striving to improve our methodologies and tools; the methodologies are rigorous and exacting which can lead to errors; and many of the tools developed through research at agencies and via federally funded initiatives use similar components (matrix estimation, shortest path algorithms, synthetic population generation, etc.). The most effective way to jump start these new tools and ensure consistency is to embrace Open Source Programming. While this has been discussed in our industry it has never been adopted in the way intended by true open source developers. Simply calling something open source is not sufficient – there are many other steps that must be taken.

In this presentation we will discuss the aspects of a successful open source program. We will cover two examples which have some, but not all, of these necessary aspects. The first is a series of strategic planning models that can be used for statewide or regional decision making: the GreenSTEP – EERPAT – SmartGAP models. GreenSTEP was developed by Brian Gregor of Oregon DOT to evaluate strategies for reducing greenhouse gas emissions in Oregon within a common modeling framework. EERPAT is FHWA’s nationally available version built upon the open source GreenSTEP that is currently being pilot tested by several state DOTs. SmartGAP, also a byproduct of GreenSTEP but with a very different purpose, was developed as part of the SHRP 2 C16 project to understand the effects of smart growth policies on travel demand.

The second example is a freight modeling framework that was developed for FHWA and CMAP in Chicago, and is now being applied in Florida. In this framework, a national model of firms and the business connections between them, which lead to freight movements, is coupled with a regional truck touring model that represents the truck movements required to deliver the shipments. In both examples, the models are a modular sequence of components that perform specific tasks, are developed using R (an open source statistics platform), and use a software repository to manage and incorporate code updates from authors at different agencies.