6.4. Implementation challenges

Concerning the implementation issues, the following three main problems arise: (1) the implementation of the fuzzy data model if fuzzy data are considered; (2) the evaluation of queries extended by fuzzy preferences; and (3) the scalability of the query evaluation, which are discussed below.

6.4.1. Modeling fuzzy databases

Existing graph database management systems make it possible to only model crisp property graph databases, so the problem of modeling fuzzy databases in such systems arises. In the case of the property graph model like that considered in Neo4j, a set of properties (key–value pairs) can be bound to a node or an edge. Properties usually denote embedded data and meta-data associated with nodes or edges. A simple mechanism may then be used to simulate fuzzy graph databases in this crisp property graph model: we only have to attach to each edge of the property graph a supplementary arbitrary property called fdegree carrying the degree...