Seed Portfolio Management with Production Flexibility

The acquisition of production flexibility is a well documented strategy pursued by many firms to counteract certain operational constraints. However these flexibilities can increase the complexity of a production system and the difficulties in managing increased complexity may hinder exploiting the full benefit of flexibility. We report one such flexibility paradox at Dow AgroSciences for an annual $800 million production decision with production flexibility, develop a novel optimization protocol to manage the flexibility, and discuss its implementation at the firm. The firm produces hybrid seeds using limited inventory of parent seeds and a production process that is subject to random variations. To handle the parent seed availability constraint and to partially mitigate the supply risk, the firm has made a significant financial investment in creating a second costly production opportunity in South America, that can be used after the first production in North America is completed. We first show that the joint problem of parent seed allocation and sequential production can be written as a tractable simultaneous problem for any realistic yield distribution. This tractable reformulation provides an exact solution in practical time durations for large assortments of several hundred hybrids. Subsequently we prove several structural results and extensions for the problem. In doing so, we make theoretical contributions to the stochastic random yield inventory literature. Our analytical solution is in use at Dow AgroSciences, with estimated savings of $15-$17 million annually. Using a numerical case study we consider two practical heuristics for allocating limited parent seed inventories that were being used in the absence of our optimization based solution. We find that in these heuristics (i) the quantities of hybrids produced are very sensitive to the sequence in which constrained capacities are allocated, (ii) parent seeds are used sub-optimally such that shortages and scarcities occur in significant magnitudes at the same time, and (iii) the heuristics completely postpone the production of some hybrids to the costlier second production incurring significant costs, which the optimal solution is able to avoid. The optimization protocol implemented at Dow avoids these pitfalls.