G. Bel et E. Bensana, ARTIFICIAL-INTELLIGENCE AND JOB-SHOP SCHE DULING SYSTEMS IN AERONAUTICAL MANUFACTURING, La Recherche aerospatiale, (1), 1996, pp. 11-22
Scheduling problems are a class of problems hard to solve for all know
n techniques. Many of research and development efforts have been condu
cted in different areas, particularly Artificial Intelligence. This pa
per presents the results of research on job shop scheduling in aeronau
tical manufacturing through the description of the OPAL software proto
type, developed at CERT. The OPAL software is dedicated to building pr
ovisional schedules and belongs to the class of knowledge-based system
s. Several types of knowledge are embedded in the system: theoretical
knowledge about scheduling constraint analysis, experimental knowledge
about priority rules used in discrete event simulation and practical
knowledge from shop floor. Solving a scheduling problem is based on th
e cooperation between these different types of knowledge. OPAL uses re
sults from fuzzy set theory, constraint based analysis and multicriter
ia aggregation techniques. The presentation is organized in two main p
arts. The first part deals with generalities about manufacturing (sect
ion I), scheduling (section II) and artificial intelligence (section I
II). The detailed description of the OPAL system is given in the secon
d part (section IV).