The development of modeling technology in an organization is typically in the hands of a small sub-group of subject matter experts. Once these models are mature enough, a natural step is the deployment of such models for more general audiences, who will in turn use these models as a tool to help execute their job responsibilities.
In the worst case, the deployment process for a model will involve training the end-user to manipulate the native tool in which the model was built (e.g. gPROMS, Aspen Plus, Dynochem, Fluent, MATLAB, Excel, etc). This training process can be challenging due to the diversity of end-users in a large corporation. This also implies the need to install the native tool on the end-users’ computers and the potential consumption of an expensive license for the rather simple “use of a model” rather than the more challenging “development of a model”.
A slightly better approach involves the development of interfaces built using event-driven languages like Visual Basic.
But both of these scenarios involves many complexities:
- the establishment of a mechanism such that the end user has access to the latest version of the model,
- the proper training of the user in the native tool,
- the sub-optimal use of expensive licenses,
- the additional need of a framework to store/archive results from the use of the models,
- the need to maintain ad-hoc interfaces, among others.