31. European Conference on Operational Research
An extended newsvendor problem formulation for setting up flexible capacity
Ralf Gössinger, Fereidoun Rashidi
We consider a risk averse manufacturer who wants to set up capacity for a production system that is faced with stochastic capacity requirements of different types. In order to cope with random fluctuations, total capacity is generated in a flexible way by multiple, non-consumable, indivisible resources. Thereby, volume and product flexibility build on the options of activating/deactivating resources as well as determining their degree of utilization. As soon as the resources differ in terms of cost, capacity level and abilities, a combinatorial problem arises. The involved decisions form a hierarchy: For setting up capacity, the basic decision on the resource configuration has to be made. Building on that, adaptation decisions on activation and utilization of individual resources contained in the chosen configuration are to be taken subject to possible requirements scenarios. For this capacity planning situation, we formulate a stochastic (scenario-based) MILP model that takes the risk attitude of the manufacturer into consideration. In addition, we inquire the structure of optimal solutions to this problem by means of a full-factorial numerical study and related statistical analyzes.