Improving Computational Efficiency of Bat-Inspired Algorithm in Optimal Structural Design

Abstract

Bat-inspired (BI) algorithm is a recent metaheuristic optimization technique that simulates echolocation behavior of bats in seeking a design space. Along the same line with almost all metaheuristics, this algorithm also entails a large number of time-consuming structural analyses in structural design optimization applications. This study is focused on improving computational efficiency of the BI algorithm in optimum structural design. The number of structural analyses required by BI algorithm in the course of design optimization is reduced considerably by incorporating an upper bound strategy (UBS) into the solution procedure. The performance of the resulting algorithm, i.e. UBS integrated BI algorithm (UBI), is evaluated in discrete sizing optimization of large-scale steel skeletal structures designed for minimum weight according to American Institute of Steel Construction-Allowable Stress Design provisions. The numerical results verify that the UBI results in a significant gain in the computational efficiency of the standard algorithm.