Successfull decision support systems depend on high quality data. These days
such data is often found in relational databases, ranging form small Access
databases to full blown ERP systems. Being able to move data reliably from those systems
into an optimization model can be an important first step in building an effective
application.
For larger data sets, a proper database can be a more reliable data store than a spreadsheet. Where spreadsheets have relative large error rates in their content, databases have tools to help protect data such as data types and referential integrity.
In many cases GDX files are useful as intermediate data stores. The attached diagram shows a complicated
data collection step where data from several sources are combined in GDX file, ready to be used
by a regional partial equilibrium model. The data sources include Microsoft Access databases, Excel
spreadsheets and GAMS include files.
GDX files can also be used as intermediate data stores within a GUI based application. This example
runs GAMS from MS Access. The data is extracted by queries and stored in a GDX file, which is read by the
GAMS model that solves a Traveling Salesman Problem (TSP) or a Minimum Spanning Tree (MST) model.
The solution is stored back in the database and displayed as a graph.
The databases for this model were converted from
MS Access and Excel files using SQL2GMS into a single large GDX file.
The combined GDX file can be used as data store for agricultural policy evaluation models.
CGE Modelers can concentrate on their GAMS models and don't need to worry further about
Access tables.
|