Computational & Technology Resources
an online resource for computational,
engineering & technology publications |
|
Civil-Comp Proceedings
ISSN 1759-3433 CCP: 106
PROCEEDINGS OF THE TWELFTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY Edited by:
Paper 174
A New Algorithm for Robust Truss Optimization: Benchmark Results A. Csébfalvi
University of Pécs, Hungary , "A New Algorithm for Robust Truss Optimization: Benchmark Results", in , (Editors), "Proceedings of the Twelfth International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 174, 2014. doi:10.4203/ccp.106.174
Keywords: optimal truss design, robust truss optimization, uncertain load direction, unified uncertainty management..
Summary
In this paper, we present a new challenging benchmark problem and its exact
solution for a recently developed robust truss optimization model when the only
source of uncertainty is the variability of the applied load directions. The essence of
the novel concept is independent from the theoretical description of the uncertainty
which may be either probabilistic (stochastic) or possibilistic (fuzzy). In the
presented unified (non-probabilistic and non-possibilistic) approach, the varying
load directions are handled by quadratically constrained parameter sets. Naturally
the presented load direction handling method can be replaced by any other
uncertainty representation which can be described by an appropriate combination of
box constraints and linear or quadratic relations. The result of the optimization is a
robust minimal-weight truss design which is invariant to the investigated load
uncertainty type. The well-known ten-bar plane truss problem with problem specific
robust additional parameters was used to illustrate the validity and efficiency of the
presented approach. The presented correctly described benchmark problem and its
exact solution may be used for testing the quality of exact and heuristic solution
procedures to be developed in the future.
purchase the full-text of this paper (price £20)
go to the previous paper |
|