Selection and scaling of spectrum-compatible ground motion records using hybrid coded genetic algorithms

This paper presents a new searching framework for optimal scaling of earthquake ground motion records as inputs for dynamics analysis. Two hybrid-coded Genetic Algorithms (GA), named real-permutation and binary-permutation GA, are effectively used to solve an applicable optimization problem in earthquake engineering field. Methodologies are outlined to choose a set of ground motions, with a good level of fit to the design spectrum, and the corresponding scales simultaneously during a hybrid-coded process. The effects of different parameters used in design of algorithms are investigated through sensitivity analysis to suggest a set of proper input values. Analysis showed that the sensitivity of the binary-permutation GA results to input parameters variations is less than that of real-permutation GA. The paper also concludes that binary-permutation GA is slightly more reliable than real-permutation GA; accordingly, it is recommended as a suitable algorithm to select and scale spectrum-compatible ground motion records.