Algorithm-aided structural-optimization strategies for the design of variable cross-section beams

The optimization of important structural components - such as beams - has always represented a crucial challenge in architectural and structural design, especially considering how the optimization of certain components often involves the loss of control of structural shapes. In this scientific contribution, an analytical model for shape optimization is presented. The shape optimization of the test-case, a variable-section beam, is modelled and optimized using two different approaches and solvers: I) MATLAB-GA®, a stochastic, population-based algorithm that randomly searches the optimal solution among population members, by mutation and crossover operators; ii) Gh-Octopus®, a Multi-Objective Evolutionary Optimization solver, which allows the production of optimized trade-off solutions between the extremes of each goal, able to support designers in decision making. The methods combine Computational Geometry and Parametric Design and allow control of the shapes of structural elements while increasing structural performance. The good accordance between the results retrieved by the numerical model implemented on MATLABGA ® with the numerical model implemented using Gh-Octopus® allowed the validation of the analytical method presented in this contribution.