Note on "Myopic Heuristics for the Random Yield Problem"

Bollapragada and Morton (1999) present several well-performing heuristics for solving the periodic inventory problem with random yield and demand. Their approach is essentially based on a transformation of the single-period problem into a standard newsvendor problem with deterministic yield and random demand which, however, is supply dependent. In our note, we show that their evaluation of the respective optimality condition is not correct. This explains the steady deterioration of their myopic heuristics for parameter constellations that correspond to increasing service levels. Some computational investigations reveal that the performance of the heuristics can become quite poor if service levels are high and exceed those values for which results are reported in the original study. Nonetheless, up to now these heuristics are still the best ones available for solving the joint random yield problem.