System reliability based multi-objective design optimization of bridges

Abstract

Continuous progress in structural reliability theory and methods, in addition to the evolving advances in optimization solution techniques, drives the steady development of reliability-based structural design process. The integration of system reliability in the design optimization of structures, especially bridges, improves the allocation of available, and usually scarce, budgets. However, the complexity in the determination of system reliability of large structural systems renders the consideration of system reliability in design optimization computationally prohibitive. This issue is even exacerbated when discrete design variables are considered. Furthermore, the optimization of structural systems typically requires using linked-discrete design variables. The recently developed matrix-based system reliability method (MSR) enables the computation of system reliability efficiently and accurately through convenient matrix calculations. The non-dominated sorting genetic algorithm with controlled elitism (NSGA-II) has proven its effectiveness and efficiency in numerous multi-objective optimization applications. The objective of this paper is to propose an approach for the system reliability-based multi-objective design optimization of bridges using the MSR method and NSGA-II and considering linked-discrete design variables. The approach is illustrated by redesigning an actual bridge in order to compare the results of the proposed approach with those of conventional bridge design.