Resilience-Based seismic design of buildings through multiobjective optimization

Abstract

The growing concern for resilience among engineers and other stakeholders is rapidly expanding the range of performance objectives imposed on buildings subjected to seismic ground motion, underscoring the need for multi-objective optimization in design. Genetic algorithms have received widespread attention in the literature as a powerful tool for multi-objective optimization of complex engineering systems, making them a prime candidate for this application. Using models of 4, 7, 10, and 15-story reinforced concrete moment frame office buildings (i.e. Risk Category II), this paper applies a resilience-based performance evaluation procedure using a genetic algorithm in order to take advantage of the potential of multi-objective optimization in designing buildings for resilience. Building models are parameterized by design variables of stiffness, strength, and deformation capacity, which are altered during optimization to enhance building performance in terms of resilience. Moreover, optimization is performed in terms of both life span performance and conditional performance at 2% and 50% probability of exceedance in 50 years and the design hazard level. In order to facilitate translation of multi-objective optimization outcomes to a final, manageable set of design alternatives to be considered by stakeholders, a pruning approach is proposed based on the existing literature and combined with the chosen genetic algorithm. Additionally, comparison of optimized Risk Category II designs with code-based Risk Category IV designs (i.e. essential facilities) facilitates assessment of the effectiveness of the more stringent stiffness and strength criteria required by building codes of essential facilities. The proposed methodology is found to be effective at identifying optimal designs given a set of constraints. The Risk Category IV criteria are generally found to be effective in terms of optimizing both loss of function and total cost.