Abstract
For the last four years, the power of Tsume-shogi solvers (i.e., programs for solving Japanese-chess problems) has advanced remarkably. Hard problems with 19 to 25 steps are now being attacked by several different types of search algorithms in the hope of solving all but a small set of these problems within a short time. In this paper, we present a parallel algorithm on network-connected distributed UNIX workstations, and show new computing results by solving 100 hard problems. The results confirm that our parallel program can solve most of the problems much more quickly than the best sequential program on a UNIX workstation.
Get full access to this article
View all access options for this article.