Smoothed

We consider various combinations of distribution estimation and resampling (bootstrap and bagging) for obtaining samples used to acquire a solution and for computing a confidence interval for an optimality gap, see paper “Distributions and Bootstrap for Data-based Stochastic Programming”.

Reference

See “Distributions and Bootstrap for Data-based Stochastic Programming” by Chen and Woodruff on the Optimization Online website.

Notes

• Both “smoothed_B_I” and “smoothed_center_sample_size” are required in the json options even though only

one of them is needed.