| dc.description.abstract | The salmon farming industry is shifting from an entrepreneurial spirit towards industrialization and cost efficiency. Large players are purchasing the small players,and the industry is consolidating. Optimization is therefore becoming more relevant,but little academic work has been done on the freshwater part of the valuechain.This thesis presents a linear stochastic programming model that minimizes the totalexpected costs related to smolt production. One of its innovations is a linearformulation using SOS2, enabling water temperature to be a decision variable. Themain uncertainty, freshwater intake temperature, is modeled through the use of scenarios,which were generated using a seasonal AR(1)-model. Testing and feedbacksessions with biologists were conducted to ensure model quality.The model was applied to two case studies involving the Marine Harvest freshwaterfacility at Slørdal, Sør-Trøndelag. The first case investigates a typical two year production plan, while the other explores the production of 500 grams smolts, whichis a new product. These case studies have yielded three core insights. First, theyindicate that optimization can have a significant impact on the total costs. A typicaltwo year production plan at the Slørdal facility experienced a total cost reductionof 11%. Second, smolt orders can be fulfilled with reduced water heating comparedto today?s praxis if the fry are deployed earlier. Third, 500 grams smolt should bedelivered during the winter to exploit the natural temperature seasonalities. Smoltdelivered during the winter has significantly lower cost and production time comparedto other seasons. | nb_NO |
