論文 - 山森 一人
-
岩切 淳一, 山森 一人, 吉原 郁夫, 相川 勝
電気学会研究会資料. IS, 情報システム研究会 2007 ( 10 ) 25 - 30 2007年4月
記述言語：日本語 掲載種別：研究論文（学術雑誌）
-
小泉 尚己, 飛高 壮明, 吉原 郁夫, 山森 一人, 安永 守利
電気学会研究会資料. IS, 情報システム研究会 2007 ( 1 ) 5 - 9 2007年4月
記述言語：日本語 掲載種別：研究論文（学術雑誌）
-
Implementation of an effective hybrid GA for large-scale traveling salesman problems
Nguyen H., Yoshihara I., Yamamori K., Yasunaga M.
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics 37 ( 1 ) 92 - 99 2007年2月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
This correspondence describes a hybrid genetic algorithm (GA) to find high-quality solutions for the traveling salesman problem (TSP). The proposed method is based on a parallel implementation of a multipopulation steady-state GA involving local search heuristics. It uses a variant of the maximal preservative crossover and the double-bridge move mutation. An effective implementation of the Lin-Kernighan heuristic (LK) is incorporated into the method to compensate for the GA's lack of local search ability. The method is validated by comparing it with the LK-Helsgaun method (LKH), which is one of the most effective methods for the TSP. Experimental results with benchmarks having up to 316228 cities show that the proposed method works more effectively and efficiently than LKH when solving large-scale problems. Finally, the method is used together with the implementation of the iterated LK to find a new best tour (as of June 2, 2003) for a 1904711-city TSP challenge. © 2007 IEEE.
-
Implementation of an Effective Hybrid GA for Large-Scale Traveling Salesman Problems 査読あり
H.D.Nguyen, I.Yoshihara, K.Yamamori, M.Yasunaga
IEEE Trans. Systems, Man, And Cybernetics, PART B:CYBERNETICS 37 ( 1 ) 92 - 99 2007年2月
記述言語：英語 掲載種別：研究論文（学術雑誌）
-
Identification of Exon-intron Boundaries by Integration of Base-oriented Genetic Programming and Statistical Heuristics 査読あり
K.Yamamori, Y.Fujita, I.Yoshihara, M.Aikawa
Proc. Artificial Life and Robotics 2007 (AROB'07) 657 - 660 2007年1月
記述言語：英語 掲載種別：研究論文（国際会議プロシーディングス）
-
Variable-Length-Segment-Transmission-Line and its Design Guidelines 査読あり
N.Koizumi, K.Hayashi, M.Yasunaga, K.Yamamori, I.Yoshihara
Proc. Artificial Life and Robotics 2007 (AROB'07) 661 - 664 2007年1月
記述言語：英語 掲載種別：研究論文（国際会議プロシーディングス）
-
Enhancement of the Variable-Length-Transmission-Line design method for multi-point optimization
Koizumi N., Yoshihara I., Yamamori K., Yasunaga M.
2006 IEEE Congress on Evolutionary Computation, CEC 2006 142 - 148 2006年12月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：2006 IEEE Congress on Evolutionary Computation, CEC 2006
Distortion of the waveform on Printed Circuit Board (PCBs) is a serious problem in higher-frequency signals transmission. To overcome this problem, we have already proposed segmental transmission line (STL). The STL divides transmission lines into multiple segments with different line widths, those are adjusted to reshape the signal waveform by superposition of reflected waves occurring at segment boundaries. In the previous work, we reshaped the waveform at only one point in the transmission line connected with a device at a time. This paper expands the method to reshape the waveform at two points of the devices at a time. We design a DIMM (Dual In-line Memory Module) clock-line for high-speed computers using our new method, and show effectiveness of the method. © 2006 IEEE.
-
佐藤 真和, 吉原 郁夫, 山森 一人, 安永 守利
宮崎大學工學部紀要 35 263 - 268 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###The eukaryotic genome consists of two kinds of regions. One is coding region that operates to ###synthesize protein. The other is non-coding region that does not operate to synthesize protein. Most of ###DNA base sequences of eukaryotic genome are non-coding region. To find embedded information in ###non-coding region is one of the most important tasks for clarifying evolution of species. ###We analyze DNA base sequences of Dictyostelium Discoideum and ribosomal protein gene with 1/f ###noise in chaos theory.
-
Estimating Physiology Activity of Functional Foods with Self-Organizing Map
Yoshihara Ikuo, Fukushima Tamon, Yamamori kunihito
宮崎大學工學部紀要 35 243 - 246 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###As interest in health and disease increases, so-called functional foods containing good-effect on ###mind and body are in the spotlight. It is necessary but time-consuming to measure body-control materials ###for estimating physiology activity. A method to estimate activity value for classification from quantity ###of protein expression is expected, because to measure quantity of protein expression is easier than to do ###physiology activity value. For estimating physiology activity value, we employ Self-Organizing Maps ###having classification ability.
-
坂口 慶幸, 吉原 郁夫, 山森 一人, 剣持 直哉
宮崎大學工學部紀要 35 269 - 274 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###Eukaryotic genes are composed of exons and introns. The former are translated into protein, but ###the latter are not. All the functions of introns are not necessarily clear. However, introns are believed ###to have close relations to the evolution of species, therefore to analyze introns is of great importance. ###To reveal hidden properties in introns, the concept of energy is extended and the extended energy is ###estimated based on Boltzmann distribution that is equivalent to the conditional probability. To compare ###the extended energy with that of other species tells the extended energy increases as the elapse time of ###evolutionary branches. The experiments lead us to the conclusion that the extended energy is a promising ###criterion to distinguish if there exists of common characteristics of species that have a common ancestor.
-
An Efficient Job Scheduling Algorithm for Grid Environment with Communication Delay
Yamamori Kunihito, Iwakiri Junichi, Yoshihara Ikuo, Aikawa Masaru
宮崎大學工學部紀要 35 315 - 320 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###We propose two scheduling algorithm for computational grid environment. One is IJQ to improve ###general job queue scheduling algorithm. Another is ASR to reduce affect of communication delay. We ###implement a grid simulator that consists of several sites to evaluate our algorithms. The simulation result ###shows that the performance of IJQ is higher when the more jobs arrive at job queue. ASR decreased ###effects of communication delay on the computational grid environment with large communication delay.
-
A Physiology Activity Value Prediction of Functional Foods Using Bayes Classifier
Yoshihara Ikuo, Tougou Syunichi, Yamamori Kunihito
宮崎大學工學部紀要 35 247 - 250 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###Body-control function of foods has been appealed in recent years. For evaluating body-control ###function of foods, it is necessary to make experiments to measure quantity of ingredients of foods for ###estimating physiology activity values, and requires a lot of time. We give a new twist to estimate physi-###ology activity values from quantity of protein expression, because to measure quantity of protein expres-###sion is easier than to experiment for estimating physiology activity values. This research employs Bayes ###classifier which minimizes estimation error of physiology activity values.
-
中川 匠, 吉原 郁夫, 山森 一人, 安永 守利
宮崎大學工學部紀要 35 257 - 262 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###Alignment based on DP-matching is used to extract unknown feature pattern embedd ed in genome ###sequence. To extract feature pattem with base length n, complete set of similarity of c andidate pattern ###with base length N = 2, 3, ・ ・ ・ , n is required. When extending base length n, number of candidate ###pattern and execution time i ncrease exponential order O(4n+1 ). This paper propose a new faster method ###of extracting feature pattern by reusing similartiy which is calculated past step. The method enable ###extracting feature pattern with base length n = 10 to speed up as much as 9 times than conventional ###method.
-
ヒューリスティクスを併用したGPによるエクソン・イントロン境界の高精度判別
藤田 祐司, 山森 一人, 吉原 郁夫, 相川 勝
宮崎大學工學部紀要 35 309 - 314 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###Genetic Programming (GP) is one of a method to identify exon-intron boundaries in gene sequences. ###In conventional method, bases are represented by 2- or 4-bits length binary codes and these bits are inde-###pendently used as inputs to the identification model. Here we develop a method to generate exon-intron ###boundaries identify model by GP with the heuristics obtained by statistical characteristics. And our ###method can directly accept not the bit-encoded data, but the character data of bases. The heuristic is ###defined as the differences of appearance frequencies for bases around the "GT" and "AT", and they are ###abstracted from 8,000,000 base sequence data. We combined GP with heuristics to Generate identifica-###tion model for Exon-Intron boundaries.
-
塩基列の長さと重複率に基づくSVMを用いたエクソン領域抽出精度の向上法
冨成 裕輔, 山森 一人, 吉原 郁夫, 藤田 祐二
宮崎大學工學部紀要 35 303 - 308 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###In this report, we evaluate the base-encoding methods to identify Exon-Intron boundaries by Sup-###port Vector Machine. In conventional experiments, 2-bits for 4-bits binary encoding was employed to ###describe bases. We focuced on that only the euclidean distance is used in Support Vector Machine, then ###propose three-tuple encoding which has the same euclidean distance among four encoded bases with ###small dimensions. In addition, we also show a new method how to decide the shortest sequence of bases ###to identify boundaries. Experiments showed that our method could achieve good identification ratio.
-
東 祐輔, 吉浜 麻生, 吉原 郁夫, 山森 一人, 剣持 直哉
宮崎大學工學部紀要 35 275 - 280 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
ABSTRACT ###The intron that does not contribute to protein synthesis has many unknown points for function and ###reason for existence, etc. It is thought that elucidation those points are useful for the clarification of the ###biological evolution and the genome function, etc. We tried analyzing intron by classification of species. ###The algorithm for classification used self-organizing map (SOM) that is able to view distribution of data ###in two-dimensional map. The using data for analysis is ribosomal protein gene that is thought useful for ###intron analysis. The codon frequency is calculated from each of transcriptional region and intron region, ###and classified species using SOM. In experimental result, it classified species in each of exon region and ###intron region. Therefore, it was understood that intron has unique information of species.
-
吉原 郁夫, 高島 弘明, 山森 一人, 菅原 研
宮崎大學工學部紀要 35 251 - 256 2006年8月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：宮崎大学
Abstract ###Discovering signs of the evolution concealed in the intron that exists in eukaryote's gene to clarify how ###the seed had evolved becomes a problem. Kind edge relation of the seed can be examined by comparing ###genomes of different seeds. It takes plenty of time to compare genomes in detail over the wide range. ###However, if the range for which it searches beforehand is understood, it is efficient. It aims to squeeze the ###range of the retrieval of the genome in this research. ###This paper proposes a method of image compression for the genome sequence and an index comparing ###seeds. Seeds are compared with compression rates corresponding to different seeds. The experiments ###reveal the compressibility is related to similarity between seeds.
-
グリッド環境に対応したジョブスケジューリング法に対するマイグレーションの評価
岩切淳一, 山森 一人, 吉原 郁夫, 相川 勝
情報処理学会研究報告ハイパフォーマンスコンピューティング（HPC） 2006 ( 87 ) 85 - 90 2006年7月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：一般社団法人情報処理学会
グリッド環境でのスタベーションを回避するため，従来のスケジューリング法とマイグレーションを組み合わせた手法について性能評価を行い，最も効率良くスタベーションを回避できる手法について提案する．シミュレーションの結果，従来のLJF/First-Fit(Largest Job First/First-Fit)にマイグレーションを組み合わせたスケジューリングにより，オーバヘッドの増大を防ぎつつスタベーションを回避し，計算機の利用効率，Slowdown Ratio共に向上させることができた．We combine migration and conventional job scheduling algorithms to avoid starvation in Grid Environment, and evaluate the performance of the combinations to propose the most efficient combination for starvation. Simulation results showd that the combination of conventional LJF/FF(Largest Job First/First-Fit) with migration could avoid the starvation with small overhead of migration. Furthermore，the LJF/FF with migration could improve both utilization of computational nodes and Slowdown Ratio．
-
GAによるセグメント分割配線構造のセグメント長と線幅の最適化
小泉 尚己, 吉原 郁夫, 山森 一人, 安永 守利
電子情報通信学会技術研究報告. CAS, 回路とシステム 105 ( 504 ) 25 - 30 2006年1月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：一般社団法人電子情報通信学会
プリント基板上を伝搬する信号を高周波化する際、プリント基板上信号の波形歪みが問題となっている。この問題に対し我々は、セグメントの幅をところどころ変えることにより波形整形を行うセグメント分割伝送線(STL : Segmental Transmission Line)を提案し小規模実験を行った。この方式を発展させ大規模な問題に対して適用するために、今回はセグメントの幅と長さを変化させることにより、利用できるセグメントの幅を限定した条件下で信号波形を整形する。そして、GAによる最適なセグメントの幅と長さを決める設計手法を提案する。さらに、単純化した物理モデル上で提案手法により設計した配線を作成し、その結果を検証する。
-
サポート・ベクター・マシンにおける遺伝子のエクソン・イントロン境界識別に適した塩基表現
冨成 裕輔, 藤田 祐司, 山森 一人, 吉原 郁夫
電子情報通信学会技術研究報告. CAS, 回路とシステム 105 ( 503 ) 55 - 60 2006年1月
記述言語：日本語 掲載種別：研究論文（学術雑誌） 出版者・発行元：一般社団法人電子情報通信学会
本報告では, 遺伝情報解析の有効な手段の1つであるサポート・ベクター・マシンを用いて, 遺伝子のエクソン・イントロン境界識別を行う. SVMを用いた研究の多くでは, 遺伝的プログラミングやニューラルネットワークなどによる研究と同様に各塩基のコード化には, 2変数表現, 4変数表現が用いられている. そこで, 我々はサポートベクター・マシンが学習データの識別にユークリッド距離を用いることに注目し, コード化した塩基間のユークリッド距離を均一としつつ, 変数を減らした3変数表現を提案する. さらに, 各表現でコード化された塩基列データを用い, 認識精度の比較・検討を行った. その結果, 3変数表現でも4変数表現と同様の高い識別率を得ることができた.