2024-09-062024-09-062022-08-091537-6494https://doi.org/10.1080/15376494.2022.2111735https://hdl.handle.net/11452/44351Wall thickness and the number of cells are important elements affecting crashworthiness. While wall thickness is expressed as the continuous design variable, the number of ribs and the number of inner structures are defined as the discrete design variable. Specific energy absorption is determined as the objective, peak crushing force is selected as the constraint. Surrogate models of these functions are obtained with artificial neural networks. A loop has been developed for the selection of the ideal parameters. The study revealed that the use of discrete and continuous design variables together improves crashworthiness.eninfo:eu-repo/semantics/closedAccessEnergy-absorption characteristicsCrashworthiness optimizationTheoretical predictionCrushing analysisSquare tubesPerformanceThicknessSectionsAluminumImproveDiscrete design variableMixed variableMulti-cell thin-walled tubeCrashworthiness optimizationArtificial neural networksTheoretical predictionScience & technologyTechnologyMaterials science, multidisciplinaryMechanicsMaterials science, characterization & testingMaterials science, compositesMaterials scienceArticle00084229230000150915105302410.1080/15376494.2022.2111735