Optimal utilization of production capacities in a dairy at risk

The rapidly changing market situation and the increased competition require a high degree of flexibility from the dairy at utilizing its capacities in order to combine prompt reaction to sales opportunities with a cost-effective production. The presented method for optimizing capacity utilization under a sales situation based on stochastic data enables such a flexibility. It combines the algorithm of Mixed-Integer Optimization with the representation of future risk factors via scenarios especially adapted to the dynamic structure of the decision problem, and with an optimization criterion of the Optimal Regret. Thus, it is secured that the realizable production policy yields results being only slightly worse than those obtained from optimal decisions under perfect knowledge of future data. As the implementation of the method simply requires standard software and a normal personal computer, no particular technical problems will arise. In comparison with approximation methods, the above mentioned method yielded improved results of approx. 0.4 Pf/kg per processed milk.