日本経営工学会論文誌
Online ISSN : 2187-9079
Print ISSN : 1342-2618
ISSN-L : 1342-2618
An Efficient Branch and Bound Algorithm for the Optimal Arrangement Problem in the Linear Consecutive-k-out-of-r-from-n:F System(<Special English Issue>-Information and Operations Management)
Hisashi YamamotoTomoaki AkibaWon Young Yun
著者情報
ジャーナル フリー

2006 年 56 巻 6 号 p. 429-438

詳細
抄録
A linear consecutive-k-out-of-n:F system consists of n components in a line. The system fails if and only if k or more consecutive components fail. A great deal of research work in this topic has been conducted since the beginning of the 1980s. The consecutive-k-out-of-r-from-n:F system is an extended system of a linear consecutive-k-out-of-n:F system. This system similarly consists of n linearly ordered components. The system fails if and only if there are at least k failed components among any r consecutive components. So, this system can represent quality control problems and inspection procedures, radar detection problems, and so on. The optimal arrangement problem in the linear consecutive-k-out-of-r-from-n:F system is to obtain the arrangement (optimal arrangement) that provides maximum system reliability within all the arrangements of components when all components don't necessarily have the same failure probability. In this paper, we propose an efficient algorithm based on a branch and bound method, for the optimal arrangement problem in the linear consecutive-k-out-of-r-from-n:F system. Our proposed algorithm conducts the following procedures (1) Searches only the arrangement that satisfies the necessary conditions for optimal arrangement, (2) Obtains reliability of the system and subsystems efficiently by "Malinowski and Preuss (1995)," and (3) Removes searches for arrangements that are no useful. We executed numerical experiments for solving an optimal arrangement problem in order to evaluate the proposed algorithm. From the results of actually solving the optimal arrangement using a computer, we showed the efficiency of our proposed algorithm, although the conclusions within the range of our executed experiments. We also show combinations of component failure probabilities that the proposed algorithm can solve more efficiently.
著者関連情報
© 2006 公益社団法人 日本経営工学会
前の記事 次の記事
feedback
Top