October 2000
Intermediate to advanced
288 pages
9h 22m
English
In this chapter I have presented a systematic approach for designing CA-based scheduling algorithms. I have considered the question of constructing local neighborhoods for the CA on the basis of program graphs. I described the CA-based scheduler working either to discover the CA rules by the GA, or in the operation mode. The results of the experiments are very promising. They show that the GA is able to discover CA rules suitable for solving the scheduling problem for a given instance of the problem. The preliminary results also show that discovered rules can be used to find optimal or suboptimal solutions of other instances of the problem that are unknown in advance.
Read now
Unlock full access