DESA: A Hybrid Optimization Algorithm for High Dimensional Functions

R.C. Addawe, J.M. Addawe, E.P. Adorio, and J.C. Magadia (Philippines)


Differential evolution, simulated annealing, hybrid algo rithm, high dimensional functions, sums of Gaussians.


In this paper, we propose Differential Evolution - Simu lated Annealing (DESA), a hybrid optimization algorithm for high dimensional functions. The addition of a new strategy based on parabolic estimation to Differential Evo lution (DE) algorithm and the incorporation of the Sim ulated Annealing (SA) algorithm for a selection strategy makes DESA a robust optimization algorithm. The pro posed hybrid algorithm is expected to obtain an acceptable solution particularly for some high dimensional functions. Experiments on the parameter estimation of sums of Gaus sians showed that the parabolic DESA algorithm outper formed the parent DE algorithm. In our investigation with two strategies: standard and parabolic and three selection modes: standard, probabilistic and incremental, we found that the parabolic DESA has shown reliability in ļ¬nding global minimum of the reference problem sets.

Important Links:

Go Back