Risk-averse + location-inventory
Noyan \cite{Noyan2012} argued about drawbacks of using pure two stage stochastic programming in cases that not happen repeatedly, like disasters. She develop \cite{RawlsTurnquist2010} model by incorporating risk-averse component, allowing decision maker to evaluate different solutions by varying risk coefficient. considering expectation as the preference criterion, the model employ conditional-value-at-risk (CVaR) as the risk measure. since integer variables are not considered in the second stage problem, she utilize the generic Benders decomposition scheme with both single-cut and multicut approaches. The proposed model evaluated on several randomly generated instances and the results are compared with value of perfect information (VPI) and the value of the stochastic solution (VSS) stochastic measures. She also apply her methodology on data of \cite{RawlsTurnquist2010} case study and elaborate the difference of solutions.