In order to optimize the tuning of Tuned Inerter Dampers (TID) in base-isolated multi-story buildings under near-fault pulse-like ground motions, this study presents a novel intelligent hybrid optimization framework that combines a Genetic Algorithm–Particle Swarm Optimization (GA–PSO) approach with a physics-informed feedforward neural network (FNN). This FNN-guided hybrid strategy offers adaptive, spectrum-aware TID parameters (inertance ratio, frequency ratio, and damping ratio) as explicit functions of the mass ratio µ, achieving faster convergence and superior performance in non-stationary pulse-dominated excitations compared to single metaheuristic techniques or traditional analytical H 2 methods (limited to stationary assumptions). Using a curated ensemble of near-fault records from the NGA-West2 database, nonlinear time-history analyses on benchmark structures that are five, ten, and fifteen stories show that, in intense pulse-like events, the pulse-optimized TID produces mean reductions of up to 25% in RMS base displacement, 22% in peak base displacement, and 20% in peak floor accelerations when compared to conventional designs. The method minimizes performance loss while maintaining strong control during far-fault and non-pulse near-fault motions. These findings demonstrate how the suggested intelligent hybrid GA–PSO optimized TID can be used more effectively and practically to increase seismic resilience in base-isolated structures situated in high-seismicity near-fault zones.
Li et al. (Fri,) studied this question.