Multiobjective Optimization under Uncertainty in the Design of Reinforced Concrete Reservoirs

The design of reinforced concrete reservoirs requires a balance between cost reduction, crack control and safety in service. This paper proposes a computational model that formulates sizing as a multiobjective optimization problem under uncertainty. The approach integrates the Finite Element Method (FEM), the FORM method and Monte Carlo simulations to estimate the reliability associated with the cracking limit state. Thickness and reinforcement variables are optimized to minimize cost and cracking, respecting regulatory and probabilistic constraints. Algorithms such as PSO, GA, and NSGA-II conduct the search. The results indicate that the uncertainty of the useful height of the section exerts a greater influence on the probability of failure in service than the strength of the materials. The model identified solutions with a cost up to 18% lower than the reference project, and the Pareto Front suggested efficient compromise alternatives, providing quantitative support for structural decision-making.