Probabilistic Performance-Based Optimal Design of Steel Moment-Resisting Frames. I: Formulation

Abstract

Significant progress has been made in the preceding two decades in the area of seismic engineering. Design codes are very quickly migrating from prescriptive procedures intended to preserve life safety to reliability-based design with less prescription intended to quantify risk associated with designs. Therefore, all stakeholders are given the opportunity to speak a common language probability and risk leading to structural designs that not only reliably preserve life safety after rare ground motions, but minimize damage after more frequent ground motions and thereby minimize life-cycle costs. Probabilistic performance-based design is in between traditional prescrip- tive design methods and full reliability-based design methodologies. The present paper provides an overview of a state-of-the-art model- code performance-based design methodology and casts this design procedure into multiple-objective optimization problems for single- story and multistory structural steel frameworks with fully and partially restrained connections. A methodology for applying an evolutionary genetic algorithm with radial fitness and balanced fitness functions is discussed in detail. A companion paper provides applications of the automated design algorithm to single-story frames and multistory frames with a variety of connection characteristics and beam-to-column moment capacity ratios. DOI: 10.1061/ASCE0733-94452007133:6757 CE Database subject headings: Seismic design; Nonlinear analysis; Connections, semi-rigid; Evolutionary computation; Optimization; Algorithms; Probabilistic methods; Inelastic action.