Performance enhancement of evolutionary search for structural topology optimisation
Finite Elements in Analysis and Design
A fully adaptive topology optimization algorithm with goal-oriented error control
Computers and Structures
A hybrid topology optimization methodology combining simulated annealing and SIMP
Computers and Structures
Hybrid population-based incremental learning using real codes
LION'05 Proceedings of the 5th international conference on Learning and Intelligent Optimization
| Hi-index | 0.00 |
The use of evolutionary algorithms for structural topology optimisation is said to be ineffective due to a considerably large number of topological design variables. However, such a problem can be alleviated by using additional numerical techniques. This paper presents the applications of simulated annealing (SA) for solving structural topology optimisation. The numerical technique termed multiresolution design variables (MRDV) is proposed as a numerical tool to enhance the searching performance of SA when dealing with topology optimisation. The approximate density distribution (ADD) and chromosome repairing techniques are employed to deal with checkerboard patterns and multiple disconnected areas on the structural topologies. The SA strategies using various sets of MRDV are implemented to solve a number of structural topology optimisation problems. The results obtained from the various optimisation strategies are illustrated and compared. The effect of using many resolutions of design variables on SA's searching performance is investigated. It is shown that the technique of MRDV is a powerful tool for the performance enhancement of SA when solving structural topology optimisation. The structural topologies obtained from employing the presented approach are comparable to those obtained from the classical gradient-based method.