Proposed a new problem solving architecture for knowledge-based integration of simulation and scheduling and describes two knowledge-based systems for railway scheduling using it: DIAPLAN and ESTRAC-III. For railway scheduling, the most important point in the problem solving processes of experts is that they integrate simulation with scheduling in an intelligent manner. To emulate these processes on a computer, the architecture consists of four major components-partial simulation, basic commands, tactical knowledge and strategic knowledge. The partial simulation and basic commands are used for simulation of train movements in a subsystem, and tactical knowledge is used for local scheduling. Strategic knowledge is used to emulate the reasoning processes by performing partial simulation and local scheduling repeatedly in a specific order. Since these components are defined independently and layered hierarchically in the architecture, one can develop a knowledge-based system in a step-by-step manner by defining the four components in the order given above.<<ETX>>
[1]
Mark S. Fox,et al.
AI and expert system myths, legends, and facts
,
1990,
IEEE Expert.
[2]
Shimji Araya,et al.
ESTRAC-II: An Expert System for Train Traffic Control in Disturbed Situations
,
1984,
ECAI.
[3]
P. Levine,et al.
Railcar distribution at the French railways
,
1990,
IEEE Expert.
[4]
S. Araya,et al.
An optimal rescheduling for online train traffic control in disturbed situations
,
1983,
The 22nd IEEE Conference on Decision and Control.
[5]
K. G. Kempf,et al.
Manufacturing planning and scheduling: where we are and where we need to be
,
1989,
[1989] Proceedings. The Fifth Conference on Artificial Intelligence Applications.
[6]
K Komaya,et al.
AN EXPERT SYSTEM FOR TRAIN-TRAFFIC CONTROL
,
1988
.
[7]
T. Fukuda,et al.
ESTRAC-III: An Expert System for Train Traffic Control in Disturbed Situations
,
1990
.