Transforming algebraic datastructures into relational datastructures in DSS

Many builders of DSS have probably experienced that they have to be able to use internal databases as well as data from external databases. We call the internal databases master- and scenario databases. Master databases are used to store data as a common source for all decision scenarios. A scenario database is used to store all data that are pertinent to a specific decision scenario. All databases may have different (types of) datamodels. This leads to a language problem for the user who has to use them. For the model part of a DSS we encounter a similar problem. In a Linear Programming model description we find an algebraic description of the necessary in- and output variables. If we have to create an interface for this model to the databases mentioned before we have another translation problem. For applications in which LP-models are used we have solved the translation problems in two ways. In one DSS we have transformed the algebraic data description into a so-called algebraic data model. This type of datamodel has been implemented for all components of the DSS. In a second DSS, with only relational data bases in it, another approach has been adopted. In model formulations the algebraic datamodel with statements specifying the relationship between algebraic datastructures and relations of the relational database has been implemented. The system takes care of the translation from one datamodel into the other. In this paper we will deal with the language problem just specified and the solutions implemented sofar.