Probabilistic performance based design multi-objective optimization for steel structures

Multi-objective optimization for the probabilistic seismic performance based design of an example moment frame steel structure is presented. Direct economic and social losses associated with seismic events, which are of interest in the current recommended frameworks for the performance based design of structures, are considered in the optimization problem defined. Three optimization objectives are selected: the initial construction cost, modeled as the weight of the structural system; expected annual economic loss associated with damage resulting from seismic hazard; and expected annual social loss resulting from seismic hazard induced damage. Hazus recommended procedures are applied in the economic and social loss calculations which include the fragility functions used in the damage analyses and injury event models implemented in the social loss calculations. The multi-objective optimization method uses a non-dominated sorting genetic algorithm strategy. The optimization results for the multiple objectives are presented and discussed in the form of Pareto fronts. Engineering demand parameters implemented for the seismic loss analysis are inter-story drifts and peak floor accelerations and are obtained using inelastic time history analysis for the ground motions associated with various seismic hazard levels. To illustrate the design procedure, loss parameters are calculated for an example steel structure located in Los Angeles, CA