[3F1-GS-10] AI application

[3F1-GS-10-01] School District Reorganization Problem with Limited Number of Simultaneous Change

〇Hitoshi Shimizu¹, Hirohiko Suwa², Tsubasa Nakamura², Tomoharu Iwata¹, Akinori Fujino¹, Hiroshi Sawada¹, Keiichi Yasumoto² (1. NTT Communication Science Laboratories, 2. Nara Institute of Science and Technology)

Keywords:School District Optimization, Integer Linear Programming, School Family

In Japan, where the population is declining, elementary and junior high schools are being consolidated and eliminated.
However, reorganizing school districts requires explanation and coordination with residents, and there is a limit to the number of school districts that can be changed simultaneously by a local government.
So far, the school district reorganization plan has been formulated as an integer linear programming problem, and a method for obtaining the optimal solution has been developed.
For this formulation, it is easy to add constraints on the number of schools to be eliminated, the number of elementary schools, or the number of city blocks in which school transfers occur.
However, to proceed with reorganizing school districts, the correspondence between elementary and junior high schools is not fixed. It is difficult to specify the number of school districts as "the city blocks where students attend elementary schools that promote to a particular junior high school."
Therefore, this study proposes a method to introduce such complex constraints into the school district reorganization planning problem.
We confirm the effectiveness of the proposed method through evaluation experiments.