Deductive databases for transport engineering

Transport data are typical of many application areas in that they arise from a variety of sources and are used in various ways. Furthermore, important information that is required in practical applications of transport databases is often not stored explicitly, but rather has to be deduced from some that is. It is therefore natural to consider the application in this field of deductive database management systems (DBMSs). These extend traditional DBMSs by permitting the definition of inference rules, default rules, complex data structures and integrity constraints, each of which can be used to provide facilities of substantial practical value. The work described here arose from the application of a novel deductive DBMS called PFL (Persistent Functional Language) to the storage and manipulation of road accident data. Although the primary entity type in such a database is that of an accident, queries are typically based upon the concept of a site. Because site information is not stored explicitly in the database, it must be deduced from auxiliary information that provides some indication of location. The combination of large amounts of data and computationally intensive queries presents extraordinary demands for the database system and has led to the development of new software techniques of high efficiency in both computation and data manipulation.