Project Description

The adoption of intelligent optimization strategies for automating the design process of civil structures is projected to grow significantly in the next decade, particularly for large and complex structural systems. Such systems comprise thousands of structural members that must be designed to satisfy multiple code- based strength and serviceability requirements. Additionally, it is often desirable to seek the lightest (i.e., optimum) structural members to minimize the overall weight of the structure; which typically renders substantial savings in construction cost and material use. For large-scale civil structures, attaining such optimum solution requires advanced computational optimization tools that transcend traditional trial-and- error design approaches. The emergence of high-level optimization schemes in the disciplines of swarm intelligence (SI) and evolutionary algorithms (EAs) have demonstrated great potential for assisting in the design process of civil structures. This project will focus on particle swarm optimization (PSO), a population-based stochastic algorithm that mimics the collective behavior of social organisms—e.g., bird flocking and fish schooling. PSO algorithms can perform an efficient and robust exploration of the vast search space of candidate designs to find near-optimum solutions to complex (real-world) problems. The objective of this undergraduate research is to further enhance the robustness of PSO for the structural design and optimization of large-scale steel frames. The project will also leverage the capabilities of high- performance computing (HPC) to examine multiple parallelization schemes for improving PSO speed and performance when applied to large-scale civil structures.

Qualifications

Experience with MATLAB or similar programming languages (e.g., Python, C++) and Microsoft Excel.