Design Automation and Optimization of Retaining Walls : Environmental Impact and Investment Cost Optimization using Genetic Algorithm

Abstract: This thesis explored the possibilities of incorporating automation and optimization inthe design process of cantilever retaining walls. The programming language Pythonhas been used to develop a program that given certain inputs performs the necessarydesign verifications according to Eurocodes and Swedish standards. The GeneticAlgorithm (GA) was chosen as optimization algorithm, where the objectives of theoptimization were defined as minimization of investment cost (IC) and environmentalimpact (EI).Optimized solutions from the program were compared with a previously designedretaining wall in a case study. Savings ranging between 15% and 30% could beobtained depending on the restrictions that were imposed on the optimization. Resultsalso indicate that the optimization algorithm tends to output retaining walls withhigher reinforcement content when optimizing for EI, leading to thinner structuralmembers in comparison to optimizations with respect to IC. A parametric analysis wasfurthermore performed to study the influence of varying heights and concrete classeson the optimized solutions.