Master of Agribusiness / Department of Agricultural Economics / Jason Bergtold / ABSTRACT Grands Moulins d'Abidjan (GMA) is a flour milling company operating in Côte d'Ivoire. It wishes to determine the optimal blend of wheat and additives that minimizes its costs of production while meeting its quality specifications. Currently, the chief miller selects the mix of ingredients. The management of the company would like to dispose of a scientific tool that challenges the decisions of the chief miller. The thesis is about building and testing this tool, an optimization model. GMA blends up to six ingredients into flour: soft wheat, hard wheat, gluten, ascorbic acid and two types of enzyme mixes. Quality specifications are summarized into four flour characteristics: protein content, falling number, Alveograph W and specific volume of a baguette after four hours of fermentation. GMA blending problem is transformed into a set of equations. The relationships between ingredients and quality parameters are determined with reference to grains science and with the help of linear regression. The optimization model is implemented in Microsoft Office Excel 2010, in two versions. In the first one (LP for Linear Programming model), it is assumed that weights of additives can take any value. In the second one (ILP for Integer Linear Programming model), some technical constraints restrain the set of values that weights of additives can take. The two models are tested with Premium Solver V11.5 from Frontline Systems Inc., against four situations that actually occurred at GMA in 2011 and 2012,. The solutions provided by the model are sensible. They challenge the ones that were actually implemented. They may have helped GMA save money. The optimization model can nevertheless be improved. The choice of relevant quality parameters can be questioned. Equations that link ingredients and quality parameters, and particularly those determined with the help of linear regression, should be further researched. The optimization model should also take into account some hidden constraints such as logistics that actually influence the decision of GMA chief miller. Finally, sensitivity analyses may also be used to provide alternative solutions.

Le procédé de fabrication de mélanges (PM) consiste à déterminer les proportions optimales à mélanger d'un ensemble de composants de façon que le produit obtenu satisfasse un ensemble de spécifications sur leurs propriétés. Deux caractéristiques importantes du problème de mélange sont les bornes dures sur les propriétés du mélange et l'incertitude répandue dans le procédé. Dans ce travail, on propose une méthode pour la production de mélanges robustes en temps réel qui minimise le coût de la recette et la sur-qualité du mélange. La méthode est basée sur les techniques de l'Optimisation Robuste et sur l'hypothèse que les lois des mélange sont linéaires. On exploite les polytopes sous-jacents pour mesurer, visualiser et caractériser l'infaisabilité du PM et on analyse la modification des bornes sur les composants pour guider le procédé vers le ``meilleur`` mélange robuste. On propose un ensemble d'indicateurs et de visualisations en vue d'offrir une aide à la décision. / The oil blending process (BP) consists in determining the optimal proportions to blend from a set of available components such that the final product fulfills a set of specifications on their properties. Two important characteristics of the blending problem are the hard bounds on the blend's properties and the uncertainty pervading the process. In this work, a real-time optimization method is proposed for producing robust blends while minimizing the blend quality giveaway and the recipe's cost. The method is based on the Robust Optimization techniques and under the assumption that the components properties blend linearly. The blending intrinsic polytopes are exploited in order to measure, visualize and characterize the infeasibility of the BP. A fine analysis of the components bounds modifications is conducted to guide the process towards the ``best`` robust blend. A set of indices and visualizations provide a helpful support for the decision maker.