Application of pool-based active learning in reducing the number of required response history analyses

Abstract

A step by step method is presented for reducing the need for a large number of response history analyses (RHAs) in developing surrogates to predict the structural responses. These surrogates, which map ground motions features and characteristics of the structural systems into structural responses, are used in deriving fragility curves; and mostly are developed using machine learning algorithms. A machine learning algorithm, depending on the complexity of the model, requires a sufficient amount of training data to predict the outputs accurately. For complicated structural models, generating training data can be computationally demanding.

Therefore, there is a need to generate the least amount of training data while preserving the accuracy of the prediction models. Towards this goal, a pool-based query-by-committee active learning (AL) algorithm is applied to choose probably the most informative data samples. The proposed methodology significantly reduces the computational cost of developing accurate surrogate models for structural response prediction.

Key Contributions

• Computational Efficiency: Significant reduction in the number of required response history analyses
• Active Learning Framework: Novel application of pool-based query-by-committee active learning in structural engineering
• Surrogate Model Development: Efficient method for developing accurate surrogates for structural response prediction
• Fragility Curve Generation: Improved computational approach for seismic fragility assessment
• Machine Learning Integration: Systematic integration of active learning with structural analysis workflows

Applications and Impact

This work addresses a critical computational bottleneck in earthquake engineering and structural reliability analysis. By reducing the number of required response history analyses while maintaining prediction accuracy, the methodology enables more efficient:

• Seismic fragility curve development
• Probabilistic seismic hazard assessment
• Performance-based earthquake engineering applications
• Large-scale structural reliability studies

Methodology

The paper presents a systematic approach that combines:

• Pool-based active learning strategies
• Query-by-committee algorithms for sample selection
• Machine learning surrogate models
• Response history analysis optimization

This interdisciplinary approach bridges machine learning, structural engineering, and computational mechanics to solve practical problems in earthquake engineering.

Recommended citation: Kiani, J., Camp, C., Pezeshk, S., & Khoshnevis, N. (2020). Application of pool-based active learning in reducing the number of required response history analyses. Computers & Structures, 241, 106355. https://doi.org/10.1016/j.compstruc.2020.106355
Download Paper