Dusonchet, F.Hongler, M. O.2013-01-072013-01-072013-01-0720022-s2.0-33751304511https://infoscience.epfl.ch/handle/20.500.14299/87654We explore the scheduling rules and the hedging levels that can be obtained by using a Restless Bandit Problem formulation of a make-to-stock production. The underlying dynamics are markov chain in continuous lime and the associate reward are piecewise linear. We observe that, the use of priority indices to sub-optimally solve the Restless Bandit problem yields, on a particular example, results close to the optimal.Dynamic schedulingMake-to-stock productionPriority indicesRestless bandittext::journal::journal article::research article