Corrigendum to “Evaluation of a genetic algorithm for constrained multi-objective structural optimization in laminated glass design” [Compos. Struct. 354 (2025) 118773] (Composite Structures (2025) 354, (S0263822324009012), (10.1016/j.compstruct.2024.118773))

The authors regret that due to trivial errors in the code implementation, some of the graphs and tables in the article [1] have to be replaced with the following, though the qualitative results and conclusions remain unaffected. The authors would like to apologise for any inconvenience caused. The correct design bending strength for chemically strengthened glass at the ULS, which replaces the value of 134.38 MPa listed in Table 4 of [1], is 112.50 MPa. Table 5 shall be replaced by the new Table 5 provided below. [Table presented] Concerning Sect. 2.2, we consider a a × b rectangular laminated glass plate, with a = 3 m and b = 4.5 m. The coefficient Ψp used in the EET model is 1.6162 10−6 mm−2 and Eq. (2.8) shall be substituted by the following: [Formula presented] Figures 7–10 of [1] are substituted by the following ones. Table 8 shall be replaced by its new version.[Figure presented][Figure presented][Figure presented][Figure presented] [Table presented] The sentence in Sect. 4.1 concerning the size and the clusters in the Pareto set (end of the second paragraph and third paragraph), is replaced with the following. “The Pareto set has size [Formula presented]. However a few solutions overlap in the C-wmax graph, because they present, in practice, the same objective function values. Solutions from 1 to 7 present thinner glass plies, with strengthening treatments on the lower plies, to pass the stress criteria. This group presents the best cost-to-stiffness ratio. Solutions from 8 to 23 are made with annealed glass and PVB interlayers. Solutions from 24 to 38, with SG interlayers, are the most expensive but present the lowest maximum deflection, guaranteeing the best mechanical performance at a higher cost.” Table 9 and Figs. 13 and 14 are substituted by their new version below. [Table presented][Figure presented][Figure presented] The sentence in Sect. 4.3 (8-th paragraph at page 12) about the efficiency of the mutation probability function is replaced by “Populations of size N = 80 and N = 120 experienced, on average, a faster convergence time by respectively 18.3 % and 18.7 %”. The results in the new Table 9 confirm our original conclusion that the optimal choice is to use a population size of N ≃ 2N.