Persistent Storage Technology for Planning and Scheduling

摘要

Knowledged-based systems can provide a key information processing aid tooperational planning, scheduling and monitoring of operations. Specifically, these systems can provide key information support for current deficiencies in crisis action planning for transportation logistics. Requirements for these systems include the ability to access, manipulate, and modify the information stored in existing databases, and a high level of collaborative and cooperative processing with the other planning agents including people and software components. Within the DARPA/Rome Lab Planning Initiative (ARPI), an intelligent information services architecture has been demonstrated which integrated cooperative user interaction and integrated information location via domain/user-oriented object representations. This effort involving participants and software components developed by Lockheed Martin, USC 151 and UCA demonstrated an experimental prototype operating in real time over the internet capable of providing information satisfying user requests making transparent to the user: (1) query relaxation and reformulation despite over-specific queries and lack of data, (2) location and selection of information sources based upon multiple selection criteria, (3) transformation of low-level data source information from databases into domain and user relevant information structures, and (4) the query language utilized. Internal communications over the internet were implemented using KQML, the Knowledge Query and Manipulation Language, a DARPA-sponsored emerging language and protocol for information exchange.