Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 97
PROCEEDINGS OF THE SECOND INTERNATIONAL CONFERENCE ON SOFT COMPUTING TECHNOLOGY IN CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING
Edited by: Y. Tsompanakis, B.H.V. Topping
Paper 10

Truss Structural Optimization using Hybrid Evolutionary Algorithms

J.G. de Moura1, G.F. Moita2 and S.R. de Souza2

1Federal Institute of Education, Science and Technology, South of Minas Gerais, Passos MG, Brazil
2Federal Centre for Technological Education of Minas Gerais, Belo Horizonte MG, Brazil

Full Bibliographic Reference for this paper
J.G. de Moura, G.F. Moita, S.R. de Souza, "Truss Structural Optimization using Hybrid Evolutionary Algorithms", in Y. Tsompanakis, B.H.V. Topping, (Editors), "Proceedings of the Second International Conference on Soft Computing Technology in Civil, Structural and Environmental Engineering", Civil-Comp Press, Stirlingshire, UK, Paper 10, 2011. doi:10.4203/ccp.97.10
Keywords: structural optimization, trusses, hybrid evolutionary algorithm, genetic algorithm, local search methods.

Summary
The use of truss structures in different architectural concepts is an everyday reality and allows for designers and engineers originality to be applied almost without restraints. Numerical structural optimization is usually a requirement for designers and engineers in order to minimize the costs of their projects. Hence, the total weight of the structures must be reduced; however, the restrictions imposed by the project also must be observed.

This paper is focused on demonstrating the application of computational methods, known as heuristics, in search of good solutions for plane truss structural problems. It presents a hybrid evolutionary algorithm for the structural optimization problem used for plane trusses. In this context, the current optimization problem aims to find the minimum weight of two-dimensional structures. The implemented algorithms, based upon the mechanisms of genetic algorithms using heuristic search strategies such as random descent method and variable neighbourhood search (VNS), try to find good solutions through reductions in the cross-sectional areas and of the length of each bar.

As a result of the inherent characteristics of the computational heuristics, there is a fundamental necessity for experimentation or calibration because most of the heuristics that are suitable for a given problem and might not always work well for a different class of problems. Consequently, parametric adjustments are essential and a considerable number of tests are imperative, so that better results can be obtained.

Two different approaches were tested in the current stage. The first is applied to problem with fixed coordinates and refines the initial solution. In the second, the refinement occurs after the population evolution and is applied to a problem with moveable nodal coordinates, allowing for shape optimization. The algorithms were implemented in conjunction with the finite element software INSANE [1], developed in Java. The details of the implementation of each heuristic can be found in [2].

Examples are shown that demonstrate the suitability of the use of heuristics to minimise the weight of such structures. The advantages and drawbacks of each method are presented and discussed, including the analysis of the results obtained in the different structural problems. The results are compared with those found in the literature and satisfactory values are attained.

References
1
F.T. Fonseca, R.L. Pitangueira, "An Object Oriented Class Organization for Dynamic Geometrically Non-Linear FEM Analysis", Proceedings of the Iberian Latin American Congress on Computational Methods in Engineering CMNE/CILAMCE, CD-ROM, Oporto, 2007.
2
J.G. Moura, "The Usage of Hybrid Evolutionary Algorithms in Problems of Trusses Structural Optimization", M.Sc. Dissertation, Federal Centre for Technological Education of Minas Gerais, Belo Horizonte, Brazil, 2009. (In Portuguese.)

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description