Takefuji Yoshiyasu

In December 1994, Japanese minor electorate system (single-member constituency system)for the House of Representatives was established where Japan is divided into three hundred constituencies. A single representative will be elected from each constituency. Zoning three hundred constituency was accomplished by hand calculators in Japan, although the constituency zoning is a very elaborate task because several constraints must be satisfied. This paper presents a neural computing approach for automatically zoning constituencies. Our method was examined by using 25 Tokyo constituencies. Ideally, the weight of a single vote in a certain population to elect a representative should be equal to that of the other constituencies. Based on the established rule, the ratio of the lightest weight to the heaviest weight must be within two. Our result shows that our ratio is 1.28 while the current (official) ratio is 1.47.

Since most of proposed parallel routing algorithms are parallelized algorithms of maze running or line search which were developed for sequential machines, efficient parallel processing with high processor utilization is difficult. Here, a parallel routing algorithm based on neural networks which can achieve both high degree of parallelism and utilization ratio is proposed, and the quality of the solution is presented on a sequential machine. The implementation on a parallel machine is also discussed.