Quantile Inverse Optimization: Improving Stability in Inverse Linear Programming

Although inverse linear programming (LP) has received increasing attention as a technique to identify an LP that can reproduce observed decisions are originally from complex system, the performance of objective function inferred by existing methods is often highly sensitive noise, errors, and uncertainty in underlying decision data. Inspired robust regression techniques mitigate impact noisy data on model fitting, “Quantile Inverse Optimization: Improving Stability Linear Programming,” Shahmoradi Lee propose notion stability develop optimization identities functions stable against imperfection. such consideration renders large-scale mixed-integer program, authors analyze connection between well-known biclique problems efficient exact algorithm well heuristics.