{"title":"Optimal Stationary Policy for a Repairable Item Inventory Problem","authors":"D. Cho","doi":"10.1111/J.1936-4490.2001.TB00250.X","DOIUrl":null,"url":null,"abstract":"This paper considers an infinite-horizon repairable-item inventory problem wherein the number of repairable units returned to the system in a given period is assumed to be equal to random proportions of the serviceable units issued in the current and the last M periods. We formulate the problem as an infinite-horizon Markov decision model and propose an algorithm for finding a steady-state optimal inventory policy, which minimizes the expected long-run total discounted cost. For speeding up the convergence of the algorithm we present an improved algorithm that utilizes error bounds. We illustrate the algorithms with numerical examples. \n \n \n \nResume \n \nCet essais examine un probleme d'inventaire d'articles reparables a horizon infini ou le nombre d'unite reparables renvoyes dans le systeme a l'interieur d'une periode donnee devrait etre egale a des mesures publiees ayant ete prises au hasard parmi les unites utilisables lors de la derniere periodes M et lors de celle en cours. Nous formulons le probleme comme un modele de decision a horizon infini de type Markov et proposons un algo-rythme pour une politique d'inventaire optimale station-naire, ce qui minimise le total attendu a long terme du cout escompte. Pour accelerer la convergence de l'algo-rythme, nous presentons un algorythme ameliore qui utilise les limites d'erreurs. Nous illustrons les algo-rythmes avec des exemples numeriques.","PeriodicalId":47349,"journal":{"name":"Canadian Journal of Administrative Sciences-Revue Canadienne Des Sciences De L Administration","volume":"11 1","pages":"130-143"},"PeriodicalIF":0.9000,"publicationDate":"2009-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Canadian Journal of Administrative Sciences-Revue Canadienne Des Sciences De L Administration","FirstCategoryId":"91","ListUrlMain":"https://doi.org/10.1111/J.1936-4490.2001.TB00250.X","RegionNum":4,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"BUSINESS","Score":null,"Total":0}
引用次数: 0
Abstract
This paper considers an infinite-horizon repairable-item inventory problem wherein the number of repairable units returned to the system in a given period is assumed to be equal to random proportions of the serviceable units issued in the current and the last M periods. We formulate the problem as an infinite-horizon Markov decision model and propose an algorithm for finding a steady-state optimal inventory policy, which minimizes the expected long-run total discounted cost. For speeding up the convergence of the algorithm we present an improved algorithm that utilizes error bounds. We illustrate the algorithms with numerical examples.
Resume
Cet essais examine un probleme d'inventaire d'articles reparables a horizon infini ou le nombre d'unite reparables renvoyes dans le systeme a l'interieur d'une periode donnee devrait etre egale a des mesures publiees ayant ete prises au hasard parmi les unites utilisables lors de la derniere periodes M et lors de celle en cours. Nous formulons le probleme comme un modele de decision a horizon infini de type Markov et proposons un algo-rythme pour une politique d'inventaire optimale station-naire, ce qui minimise le total attendu a long terme du cout escompte. Pour accelerer la convergence de l'algo-rythme, nous presentons un algorythme ameliore qui utilise les limites d'erreurs. Nous illustrons les algo-rythmes avec des exemples numeriques.
期刊介绍:
The Canadian Journal of Administrative Sciences (CJAS) is a multidisciplinary, peer-reviewed, international quarterly that publishes manuscripts with a strong theoretical foundation. The journal welcomes literature reviews, quantitative and qualitative studies as well as conceptual pieces. CJAS is an ISI-listed journal that publishes papers in all key disciplines of business. CJAS is a particularly suitable home for manuscripts of a crossdisciplinary nature. All papers must state in an explicit and compelling way their unique contribution to advancing theory and/or practice in the administrative sciences.