Abstract Scheduling is a difficult task. Much of the difficulty stems from the need to allocate the limited resources to attend to a large and diverse set of constraints. In this paper, the author highlights the conflict situations that occur between constraint satisfaction and resource availability and among constraints themselves. These conflict situations hamper the satisfaction of contraints. Through a detailed analysis of the scheduling process, it is suggested that the central problem in scheduling is how to predict the future impact of current decisions. In answer, two predictive techniques are proposed, reinforcement scheduling and a decomposition approach. Reinforcement scheduling advocates that before constructing a schedule some rough utilization plans (reinforcement plans) can be built as a basis for generating decision-making guidelines for detailed scheduling. The decomposition approach argues that it is ineffective to try to satisfy all the constraints in a single scheduling process. This method considers different constraints in different processes. Based on these ideas a scheduling system called RESS-II has been implemented. Test results have shown that these techniques are effective in performing their tasks.
[1]
Bing Liu,et al.
Reinforcement planning for resource allocation and constraint satisfaction
,
1988
.
[2]
Stephen F. Smith,et al.
Constructing and Maintaining Detailed Production Plans: Investigations into the Development of Knowledge-Based Factory Scheduling Systems
,
1986,
AI Mag..
[3]
Gideon Halevi.
The role of computers in manufacturing processes
,
1980
.
[4]
N. Sadeh,et al.
Preference Propagation in Temporal/Capacity Constraint Graphs
,
1989
.
[5]
Stephen F. Smith,et al.
ISIS—a knowledge‐based system for factory scheduling
,
1984
.
[6]
Bing Liu,et al.
Scheduling via reinforcement
,
1988,
Artif. Intell. Eng..
[7]
Stephen F. Smith,et al.
Viewing scheduling as an opportunistic problem-solving process
,
1988
.
[8]
Andrew Kusiak,et al.
Expert systems for planning and scheduling manufacturing systems
,
1988
.