Implementation of Eurocode load cases in optimization problems of
steel frames, based on genetic algorithms
Ioana D. Balea
1,a
, Radu Hulea
2,b
and Georgios E. Stavroulakis
3,c
1,2
Technical University of Cluj-Napoca 400020 Cluj-Napoca, Romania,
3
Technical University of Crete Kounoupidiana, Akrotiri, Greece
ioana.balea@mecon.utcluj.ro, gestavr@dpem.tuc.gr
Keywords: optimization, steel, eurocode, genetic algorithms
Abstract. This paper presents an implementation of Eurocode load cases for discrete global optimi-
zation algorithm for planar structures based on the principles of finite element methods and genetic
algorithms. The final optimal design is obtained using IPE sections chosen as feasible by the algo-
rithm, from the available steel sections from industry. The algorithm is tested on an asymmetric
planar steel frame with promising results.
Introduction
In all engineering fields, designers attempt to find solutions that combine performance and satisfac-
tion of critical requirements. But these techniques and requirements are subject to constant change.
Engineers have always been tackling design problems defined by competing goals and bounded by
technical, aesthetic and economic constraints. Designers can obtain the optimum within the imposed
conditions by using optimization techniques. In the field of structural engineering, structures de-
signed in this way are safer, more reliable and less expensive than the traditional designed ones,
where the success of the design is based solely on the experience of the engineer. Optimization
techniques also require some expertise, but with the implementation of these algorithms in comput-
er aided design software it can become a very powerful tool in the hands of engineers. In general,
the optimization techniques used in structural design can be categorized into classical and heuristic
search methods. Classical optimization methods such as linear programming, nonlinear program-
ming and optimality criteria often require substantial gradient information. In these methods the fi-
nal results depend on the initially selected points and the number of computational operations
increases as the size of the structure increases. The solution in these methods does not necessarily
correspond to the global optimum. Many engineering design problems are too complex to be han-
dled with mathematical programming methods. In comparison, heuristic search methods do not re-
quire the data as in the conventional mathematical programming and have better global search
abilities than the classical optimization algorithms [1]. For the past 60 years a new branch of opti-
mization techniques was continuously developed, which mimics the design methods existing in na-
ture. Genetic algorithms, simulated annealing and evolutionary strategies are among such
algorithms that are used in the design optimization of structures. Among these, genetic algorithms
are a search method that is based on the principal of the survival of the fittest and adaptation. They
operate on a population of design variables sets. Each population consists of individuals that are po-
tential solutions to the design problem. A fitness value is calculated for each individual using the
objective function and constraints as a measure of performance of the design variables. If the indi-
vidual is fit, it is selected as candidate to take part in the construction of the next population, if not
then it is discarded. The solution to a general optimization process can be associated with this sys-
tem behavior. The genetic algorithm (GA) is one of the best-known heuristic methods; it has been
used to solve structural optimization problems by some researches such as Rajeev and Krishna-
moorthy [2], Saka and Kameshki [3], Camp et al. [4], Pezeshk et al. [5], Erbatur et al. [6], Kaveh
and colleagues [7], Shook et al. [8], among many others.
Applied Mechanics and Materials Vol. 310 (2013) pp 609-613
Online available since 2013/Feb/27 at www.scientific.net
© (2013) Trans Tech Publications, Switzerland
doi:10.4028/www.scientific.net/AMM.310.609
All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP,
www.ttp.net. (ID: 79.130.15.225-19/03/13,07:10:16)