Research output: Contribution to Journal/Magazine › Journal article › peer-review
<mark>Journal publication date</mark> | 16/06/2006 |
---|---|
<mark>Journal</mark> | European Journal of Operational Research |
Issue number | 3 |
Volume | 171 |
Number of pages | 12 |
Pages (from-to) | 879-890 |
Publication Status | Published |
Early online date | 10/12/04 |
<mark>Original language</mark> | English |
In this paper, we propose a successive approximation heuristic which solves large stochastic mixed-integer programming problem with complete fixed recourse. We refer to this method as the Scenario Updating Method, since it solves the problem by considering only a subset of scenarios which is updated at each iteration. Only those scenarios which imply a significant change in the objective function are added. The algorithm is terminated when no such scenarios are available to enter in the current scenario subtree. Several rules to select scenarios are discussed. Bounds on heuristic solutions are computed by relaxing some of the non-anticipativity constraints. The proposed procedure is tested on a multistage stochastic batch-sizing problem.