Knowledge based optimization modeling for decision support in system KOMIS

The author presents a current experience with knowledge-based formulation of optimization models for large-scale, long-term decision problems. The project considered originated with an attempt to verify the feasibility of integration of quantitative decision tools with emerging knowledge engineering technology. A prototype integrated system, KOMIS, took the form of a data-driven, interactive program. The basic mode of operation implemented consists of the translation of `natural language´ queries into decision problem(s), formulation of corresponding mathematical optimization models, solving the models, and presenting the user with solutions in a user-friendly manner. An example of KOMIS application is presented