Risk-averse Approaches for a Two-Stage Assembly-to-Order Problem

Assembly to order is a production strategy where components are manufactured under demand uncertainty and end items are assembled only after demand is realized. Risk-neutral approaches aim to maximize the expected profit. However, this approach may fail if heavy-tailed or multi-modal distributions are likely to generate significant disruptions or if the shrinking life of products is considered. Conversely, risk-averse models may tackle these problems. In the paper, we deal with an assembly-to-order problem, modeled as a two-stage stochastic linear programming problem considering the introduction of a classical risk measure from finance: the conditional value-at-risk. We examine the characteristics and the performance of the model by means of a large number of out-of-sample scenarios.