Multi-Resolution Optimisation: Application Of Meta-Heuristic In Function Remodelling

Multi-resolution approach to optimisation involving remodelling of objective function has great received interests lately. There are two different goals of remodelling for the objective function. First, the reduction of computational complexity of the objective functions in order to accelerate the searching process. Second, the removal of local optima in the objective functions to avoid premature convergence. Most of the approach shares the problem: the setting of the smoothness pressure value over the objective function. We proposed a meta-heuristic approach to determine the suitable smoothness pressure value throughout the searching process. The results show that meta-heuristic has significant improvement compared to state-of-theart Classic Evolutionary Programming (CEP) and Fast Evolutionary Programming (FEP).