Keywords: structural optimization, discrete optimum design, stochastic search techniques, ant colony optimization method.

The area of structural optimization has undergone a new direction starting with the introduction of meta-heuristic optimization techniques. The initial success of these methods observed in various disciplines of engineering, such as computational structural mechanics, motivated further research leading to a better understanding and improvement of these methods. These ongoing endeavours, when combined with pioneering ideas for imitating different phenomena of the nature, have led to the development of a variety of promising search methods. The basic concept behind each of these techniques rests on simulating the paradigm of a biological, chemical, or social system that is automated by nature to achieve the task of optimization of its own. Ant colony optimization (ACO) is one of the meta-heuristic methods recently developed in literature. The technique proposed by Dorigo et al. [1] is inspired from the way that ant colonies find the shortest route between a food source and their nest. It has been successfully been applied to problems from discrete structural optimization literature [2,3].

This study is concerned with reformation of the ACO method for structural optimization problems that consist of many design variables and, or large discrete sets. It is observed that a standard ACO algorithm usually exhibits serious disadvantages when applied to such problems, resulting in poor convergence characteristics and inefficient search process. The so-called pheromone scaling approach is introduced in the paper to eliminate the observed drawbacks of the standard algorithm. Moreover, a new reformulation of local update parameter is proposed by implementing it dynamically during the optimization process. Two design examples are used mainly to examine and compare numerical performance of the proposed ACO algorithm. These examples are a 160-bar pyramid and a 693-bar braced barrel vault space steel truss systems. In both examples the trusses are sized for minimum weight considering a code imposed stress, stability and displacement limitations. The solutions to these examples obtained with the proposed ACO algorithm are compared with those of the standard ACO algorithms as well as of the other meta-heuristic search techniques. A comparison of optimum designs reported for these two problems with different techniques and other variants of ACO technique reveals that the proposed ACO algorithm produces improved results and is fairly effective in finding optimum solutions to large-scale structures.

1

M. Dorigo, V. Maniezzo, A. Colorni, "Distributed optimization by ant colonies", Proc. 1st European Conf. on Artificial Life, MIT Press, Cambridge, Mass., 134-142, 1991.

2

C.V. Camp, J.B. Bichon, S.P. Stovall, "Design of Steel Frames Using Ant Colony Optimization", Journal of Structural Engineering, ASCE, 131(3), 369-379, 200. doi:10.1061/(ASCE)0733-9445(2005)131:3(369)

3

A. Kaveh, S. Shojaee, "Optimal Design of Skeletal Structures Using Ant Colony Optimization", International Journal for Numerical Methods in Engineering, 70, 563-581, 2007. doi:10.1002/nme.1898

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
purchase this book (price £70 +P&P)