Parallel simulated annealing by generalized speculative computation
Document Type
Conference Proceeding
Publication Date
12-1-1993
Abstract
Simulated annealing is known to be highly sequential due to loop carried dependencies. This report presents a new approach to parallel simulated annealing, called Generalized Speculative Computation (GSC). While the conventional binary speculative computation can execute a maximum of (log n) iterations in parallel on n processors, the GSC can execute n iterations in parallel while maintaining the same decision sequence as a sequential version. To demonstrate the performance of GSC, we implement 100 to 500 city Traveling Salesman Problem on the AP1000 massively parallel machine. Actual experimental results demonstrate that the GSC is highly effective for the given problem as we obtain a nearly 20-fold speedup for the initial temperature of 1 on 100 processors.
Identifier
0027873414 (Scopus)
ISBN
[081864222X]
Publication Title
Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing
First Page
416
Last Page
419
Recommended Citation
Sohn, Andrew; Wu, Zhihong; and Jin, Xue, "Parallel simulated annealing by generalized speculative computation" (1993). Faculty Publications. 16953.
https://digitalcommons.njit.edu/fac_pubs/16953
