## 抄録

This paper presents a new algorithm for solving frequency assignment problems in cellular mobile systems using constraint satisfaction techniques. The characteristics of this algorithm are as follows: 1) instead of representing each call in a cell (a unit area in providing communication services) as a variable, we represent a cell (which has multiple calls) as a variable that has a very large domain, and determine a variable value step by step, 2) a powerful cell-ordering heuristic is introduced, 3) a branch-and-bound search that incorporates forward-checking is performed, and 4) the limited discrepancy search is introduced to improve the chance of finding a solution in a limited amount of search. Experimental evaluations using standard benchmark problems show that this algorithm can find optimal or semi-optimal solutions for these problems, and most of the obtained solutions are better than or equivalent to those of existing methods using simulated annealing, tabu search, or neural networks. These results show that state-of-the-art constraint satisfaction/optimization techniques are capable of solving realistic application problems when equipped with an appropriate problem representation and heuristics.

本文言語 | 英語 |
---|---|

ページ（範囲） | 888-894 |

ページ数 | 7 |

ジャーナル | IEEE Vehicular Technology Conference |

巻 | 2 |

出版ステータス | 出版済み - 1月 1 2000 |

外部発表 | はい |

イベント | VTC2000: 51st Vehicular Technology Conference 'Shaping History Through Mobile Technologies' - Tokyo, Jpn 継続期間: 5月 15 2000 → 5月 18 2000 |

## !!!All Science Journal Classification (ASJC) codes

- コンピュータ サイエンスの応用
- 電子工学および電気工学
- 応用数学