Author/Authors :
Kalatjari, V. R Department of Civil Engineering - Shahrood University of Technology, Shahrood , Talebpour, M. H Department of Civil Engineering - Shahrood University of Technology, Shahrood
Abstract :
In this article, by Partitioning of designing space, optimization speed is tried to be increased
by GA. To this end, designing space search is done in two steps which are global search and
local search. To achieve this goal, according to meshing in FEM, firstly, the list of sections
is divided to specific subsets. Then, intermediate member of each subset, as representative
of subset, is defined in a new list. Optimization process is started based on the new list of
sections which includes subset’s representatives (global search). After some specific
generations, range of optimum design is indicated for each designing variable. Afterwards,
the list of sections is redefined relative to previous step’s result and based on subset of
relevant variable. Finally, optimization will be continued based on the new list of sections
for each designing variable to complete the generations (local search). In this regard, effect
of dimension and number of subset’s members of global and local searches in proposal are
investigated by optimization examples of skeletal structures. Results imply on optimization
speed enhancement based on proposal in different cases proportional to simple and
advanced cases of GA.
Keywords :
optimization , skeletal structures , GA , sampling design space
