Direct Economic Loss-Based Seismic Design Using Genetic Algorithm

Abstract

Earthquake-induced economic losses in buildings are likely to escalate as urbanization accelerates. Traditional seismic optimization methods, such as consistent ductility-based design, aim to achieve uniform story drift along the building height but do not directly address reduction in economic losses. This study presents an innovative method called direct loss-based design (DLBD) using a genetic algorithm to minimize the expected annual loss (EAL) in RC frame structures conditioned on constant structural stiffness and construction cost. Two optimization frameworks for minimizing the expected losses (EL) are examined: intensity-based and risk-based. Intensity-based optimization focuses on a specific intensity level, whereas risk-based optimization considers the EAL across three seismic hazard levels. Additionally, the collapse probability (CP) of structures optimized via the DLBD method is evaluated. Results show that intensity-based optimization significantly reduces EL at the target intensity level but is ineffective at other seismic intensities due to extremely increased seismic demands at some floors. Conversely, the risk-based optimization substantially reduces both EAL and CP across all intensity levels assessed, with a maximum reduction of 31% and an average reduction of over 20% for EAL, as well as an average decrease of 62% for CP. This approach effectively decreases seismic demands while simultaneously distributing the story drifts evenly across the building height. The proposed risk-based design framework not only offers a practical algorithm for seismic optimization targeting EL but also provides new avenues to explore resilience-based seismic design approaches.