B. Dizangian , M. R Ghasemi,
Volume 5, Issue 2 (3-2015)
Abstract
A Reliability-Based Design Optimization (RBDO) framework is presented that accounts for stochastic variations in structural parameters and operating conditions. The reliability index calculation is itself an iterative process, potentially employing an optimization technique to find the shortest distance from the origin to the limit-state boundary in a standard normal space. Monte Carlo simulation (MCs) is embedded into a design optimization procedure by a modular double loop approach, which the self-adaptive version of particle swarm optimization method is introduced as an optimization technique. Double loop method has the advantage of being simple in concepts and easy to implement. First, we study the efficiency of self-adaptive PSO algorithm inorder to solve the optimization problem in reliability analysis and then compare the results with the Monte Carlo simulation. While computationally significantly more expensive than deterministic design optimization, the examples illustrate the importance of accounting for uncertainties and the need for regarding reliability-based optimization methods and also, should encourage the use of PSO as the best of evolutionary optimization methods to more such reliability-based optimization problems.
M. Salar, M. R. Ghasemi , B. Dizangian,
Volume 6, Issue 1 (1-2016)
Abstract
Due to the complex structural issues and
increasing number of design variables, a rather fast optimization algorithm to
lead to a global swift convergence history without multiple attempts may be of
major concern. Genetic Algorithm (GA) includes random numerical technique that
is inspired by nature and is used to solve optimization problems. In this
study, a novel GA method based on self-adaptive operators is presented. Results
show that this proposed method is faster than many other defined GA-based
conventional algorithms. To investigate the efficiency of the proposed method,
several famous optimization truss problems with semi-discrete variables are
studied. The results reflect the good performance of the algorithm where
relatively a less number of analyses is required for the global optimum
solution.
