Using large data bases for interactive problem solving

Abstract

Interactive problem solving involves user-machine dialogues to solve unstructured problems, such as selecting marketing strategies and planning mass transit routes. A characteristic of most interactive problem solving systems is that they operate on data bases that are special purpose subsets derived from a large data base. The problem solving system must include code for interfacing with this data base, or the data must be converted to the formats required by the problem solving system. Effective interactive problem solving requires a data management system which provides flexibility in accessing a large data base and fast performance for the problem solving system. This paper describes a method, called data extraction, for interfacing large data bases with interactive problem solving systems. Figure 1 shows the basic components in the interface. A large data base management system is used to maintain and protect a set of data files. Data extraction is used to aggregate and subset from this set to provide information for problem solving. In addition to an I/O interface, data extraction provides interactive data description and presentation functions.