Skip to main content
placeholder image

Truss topology optimisation using genetic algorithms

Conference Paper


Abstract


  • The 10 bar truss problem has been used by many researchers to test their optimisation algorithms. This paper investigates the application of genetic

    algorithms (GAs) to optimising redundant trusses allowing variable topologies and configurations. The genetic algorithm (GA) optimisation was carried out using a qary scheme and floating point representation. Unstable, infeasible truss candidates were identified by conducting a node to node stability check. The probability of a structural member being absent is shown to have a marked effect on the location of the global optimum when considering a member size and topology optimisation. With low probabilities of null members, the GA did not find the global optimum topology. For higher probabilities, the optimal six-bar topology is found. When configuration is introduced as a design variable in addition to member size and topology, the probability of a member being null seems to have less (if any) impact. The paper concludes that the GA approach has good success in solving the 10 bar truss problem with variable member size and topology. When applied to the problem with configuration introduced as an additional design variable, the GA was able to identify potential optimal solutions which would probably not have been considered when using alternative optimisation techniques.

Publication Date


  • 2007

Citation


  • McCarthy, T. J. & Fenwick, A. G. (2007). Truss topology optimisation using genetic algorithms. In B. HV. Topping (Eds.), Ninth International Conference on the Application of Artificial Intelligence to Civil, Structural and Environmental Engineering (p. [18]). Stirlingshire, Scotland: Civil-Comp Press.

Ro Metadata Url


  • http://ro.uow.edu.au/engpapers/3077

Start Page


  • [18]

Place Of Publication


  • http://www.civil-comp.com

Abstract


  • The 10 bar truss problem has been used by many researchers to test their optimisation algorithms. This paper investigates the application of genetic

    algorithms (GAs) to optimising redundant trusses allowing variable topologies and configurations. The genetic algorithm (GA) optimisation was carried out using a qary scheme and floating point representation. Unstable, infeasible truss candidates were identified by conducting a node to node stability check. The probability of a structural member being absent is shown to have a marked effect on the location of the global optimum when considering a member size and topology optimisation. With low probabilities of null members, the GA did not find the global optimum topology. For higher probabilities, the optimal six-bar topology is found. When configuration is introduced as a design variable in addition to member size and topology, the probability of a member being null seems to have less (if any) impact. The paper concludes that the GA approach has good success in solving the 10 bar truss problem with variable member size and topology. When applied to the problem with configuration introduced as an additional design variable, the GA was able to identify potential optimal solutions which would probably not have been considered when using alternative optimisation techniques.

Publication Date


  • 2007

Citation


  • McCarthy, T. J. & Fenwick, A. G. (2007). Truss topology optimisation using genetic algorithms. In B. HV. Topping (Eds.), Ninth International Conference on the Application of Artificial Intelligence to Civil, Structural and Environmental Engineering (p. [18]). Stirlingshire, Scotland: Civil-Comp Press.

Ro Metadata Url


  • http://ro.uow.edu.au/engpapers/3077

Start Page


  • [18]

Place Of Publication


  • http://www.civil-comp.com