A unified approach to parameter selection in meta-heuristic algorithms for layout optimization

Meta-heuristic optimization algorithms have attracted many researchers in the last decade. Adjustment of different parameters of these algorithms is usually a time consuming task which is mostly done by a trial and error approach. In this study an index, namely convergence factor (CF), is introduced that can show the performance of these algorithms. CF of an algorithm provides an estimate of the suitability of the parameters being set and can also enforce the algorithm to adjust its parameters automatically according to a pre-defined CF.In this study GA, ACO, PSO and BB–BC algorithms are used for layout (topology plus sizing) optimization of steel braced frames. Numerical examples show these algorithms have some similarities in common that should be taken into account in solving optimization problems.

Research highlights
► A unified parameter adjustment for different meta-heuristic algorithms.
► GA, ACO, PSO and BB-BC algorithms for layout optimization.
► Optimal design of steel braced planar frames.
► A method to enforce the algorithm to adjust its parameters in an adaptive manner.