Concrete is widely used in structures for its durability, versatility, and strength.There is currently a push in the industry to minimize its use, driven by cost efficiencyand the need to lower carbon emissions. Deep beams are structural elementsused when there is a need for high load-bearing capacity and resistance tohigh vertical loads on relatively short spans. They are common in cantilever wallswith a room-separating function and openings. Due to the non-linear behavior ofdeep beams, classic beam theory does not apply, and they are often designed withthe Strut and Tie method, which gets increasingly more complex when openingsare introduced in the structure. Furthermore, due to the non-linearity of deepbeams, non-linear FEM is suitable for analysis.Parametric modeling is a process that allows the shape of a model’s geometry tochange automatically when dimension values are modified. This is achieved usingdesign computer programming code to define the dimensions and shape of themodel. Thus, parametric modeling can be used to simplify the design processand provide possibilities to make quick design alterations, making it suitable forautomation tasks. Genetic algorithms can be used as an optimization tool formaterial usage in a structure, regarding carbon dioxide emissions. In this thesis,Grasshopper with the plug-in Karamba3D, is used for parametric modeling andautomation. The component Galapagos is used in Grasshopper as an optimizationalgorithm, and Atena Engineering 2D for result evaluation through non-linear FEanalysis.The thesis acts as a proposal for a work method with automation of reinforcementlayout in a concrete deep beam and optimization regarding carbon dioxideemissions. Objectives include creating a universal script that automates the reinforcementlayout based on linear elastic FEM for any given geometry of a deepbeam with an opening. Other objectives refer to the formulation of a fitness function,the development of an effective optimization component, and the utilizationof non-linear FEM analysis to evaluate results and refine the script.The resulting script offers automation of reinforcement layout that effectively ensuressufficient reinforcement in models with two main limitations: oversight ofstress redistribution and high mesh dependency in the linear elastic analysis maylead to inaccuracies on which the reinforcement layout is based. Users can optimizeCO2-e levels, but informed decisions are crucial. Galapagos optimizationand Karamba settings require an understanding of the fitness function and geneticalgorithm techniques. Finally, it is recommended that the thesis is used as a basisfor a work methodology when using the script.