Efficient selection of up to three-component ground excitation sets for earthquake engineering applications

Ground motion record selection methodologies are commonly developed to ensure that the input excitation used in response history analyses embodies essential conditions such as spectral compatibility, hazard and intensity measure consistency, seismological site-specific criteria, always performing a computationally efficient manner. A methodology utilizing genetic algorithms is revisited here, expanded select multi-component ground motions satisfying typically required objectives of earthquake engineering applications, while ensuring increased efficiency. Multi-objective optimization performed, claimed be superior delivering robust results account for compatibility first second order statistics (mean standard deviation) wide range values, well criteria. unique contribution ability include probability distribution targets specific ordinates spectrum, on top mean deviation, allowing more refined sets can reduce number records analyses. Additionally, novel benchmarking process assess efficiency introduced terms providing globally-optimal solutions problem. Through this algorithm, proposed appears impeccable extracting best possible sets.