Utility network derivation from legacy source data for feature-based AM/FM systems
This paper presents a deductive inference approach for populating target Automated Mapping/Facilities Management (AM/FM) platforms with utility network data from primitive, unstructured source data. The approach exploits a declarative rulebase, which has been developed in PROLOG, to specify feature-creation constraints and to infer relationships, including the derivation of networks, semantic relation instances, and intrinsic associativity among spatial and thematic components, in legacy source data. The approach is novel in that it provides a flexible, integration framework capable of inferring utility network relationships in legacy data originating from a variety of sources. Although the approach is experimental, it has been successfully utilized in industrial applications in which AM/FM network models of electrical-distribution and broadbanddistribution systems were derived from legacy data. In all instances, the source data contained only a subset of the inter- and intra-feature relationship instances that were modelled in the target AM/FM platform.