Computers and Industrial Engineering 46 (3), 431–441]. The solution algorithm proposed there, is partly heuristic and so it leads to an approximate policy. Computational experience shows that this algorithm behaves very well and gives, in many cases, results close to the optimum. The optimization algorithm proposed in this article leads to the exact optimal policy within a specific class of policies. These academic papers are available to purchase through Sciencedirect.com, usually at US$30 each. To do this, it is necessary to register via the weblink given.

External Link