Expert Tower Analysis and Design System. II: Search Strategies and Learning

Abstract

This paper deals with the various search strategies for generation of alternate feasible designs that have been adopted in EXTASY—an expert system for analysis and design of microwave towers. Heuristics guide the search by varying certain metageometry parameters. The IF‐THEN‐WITH format of knowledge representation is used to make the system “learn” new search strategies as it solves more problems. The search strategies evolve as the system solves more design problems. The search will be guided by the data abstracted from a previously designed tower, whose input vector differs least from the tower to be designed. A sample session with EXTASY is also included. The system is implemented on an IBM PC‐AT with MS‐DOS. The languages used are PROLOG, FORTRAN and BASIC.