GARI
GARI is probably the first AI based CAPP system to be reported in the literature, developed by Descotte and Latcombe (1981) at the University of Greonoble in France. GARI consists of a knowledge base of about 50 rules and a general purpose problem solver. Knowledge is represented by production rules dealing with conditions for the part being manufactured and advises representing technological and economical preferences. These preferences are weighted according to their importance and, where pieces of advices conflict, the system refers to their weights. The manufacturing rules are of the 'IF-THEN' type and are parameterized by simple variables and set variables (see Figure 5). GARI is implemented in MACLISP language and operates on CH-Honeywell Bull HB-68 computer under the MULTICS operation system. Recently GARI has been tested in the metal cutting industry with satisfactory results. A typical GARI representation is shown in Figure 6.
Figure 5: Rule Based System
Figure 6: A Typical GARI Representation
(F1 (type face) (direction xp) (quality 120))
(F2 (type face) (direction yp) (quality 64))
(F3 (type face) (direction ym) (quality rough))
(H1 (type countersunk- hole) (diameter 1. 0)
(Countersink- diameter 3.0)
(Starting- from F2) (Opening- into F3))
(Distance H1 F1 3.0)
(Countersink- depth F2 H1 0. 5)