2004-09-13Buch DOI: 10.18452/8325
On deviation measures in stochastic integer programming
We propose extensions of traditional expectation-based stochastic integer programs to mean-risk models. Risk is measured by expected deviations of suitable random variables from their means or from preselected targets. We derive structural properties of the resulting stochastic programs and present first algorithmic ideas to achieve problem decomposition.
Dateien zu dieser Publikation