Abstract: This paper introduces an expert system in the real-time control for the after-blow operation in converter steelmaking with oxygen. Special attention is paid to the research on the representation structure of the knowl edge-base and the relative inference algorithm.
A two-time mapping model for the solution is formed. An improved opportunity inference approach in the problem is employed, instructed by the means-end analysis with heuristic information and metarules. In this way, the production system has parallel processing performance, and the expert system has a high adaptability.

Key words: solving strategy, expert system, knowledge-base