Optimum shape of large-span trusses according to AISC-LRFD using Ranked Particles Optimization

摘要

The purpose of this study is the application of the recently developed metaheuristic algorithm named Ranked Particles Optimization (RPO) in shape optimization of large-span truss structures and comparing it with Particle Swarm Optimization (PSO) and Harmony Search (HS) as two traditional and widely-used methods. The study considers the design of large-span truss structures mimicking real-world application of metaheuristics in the design of civil engineering infrastructures. Since the results metaheuristic algorithms depend on random initial variables, the design codes are run several times to obtain a sample pool. Then, statistical analyses are performed to compare the performance of each algorithm within the considered class of problems. The results were presented to study the convergence rate, robustness in finding the global minimum, and the repeatability of the RPO. The results exhibit that RPO converges faster compared to the PSO and the HS algorithms. Moreover, the results of PRO have less cost function which is an indication of the robustness of the algorithm. Finally, less dispersity of the results of RPO compared to the PSO and HS shows high repeatability and reliability of the algorithm. (C) 2017 Elsevier Ltd. All rights reserved.