MVMO for bound constrained single-objective computationally expensive numerical optimization

Mean-Variance Mapping Optimization (MVMO) is a recent addition to the heuristic optimization field. The main traits of its evolutionary mechanism reside in the adoption of a single parent-offspring pair approach along with a normalized range of the search space for all optimization variables as well as in the use of a special mapping function, which accounts for the actual mean and variance of the normalized optimization variables for mutation operation. MVMO is also open to further extensions and hybridization with other approaches. Along this spirit, this paper surveys the performance of MVMO when executed in its pure algorithmic procedure and when hybridized to include a local search strategy. Numerical experiments are conducted on the IEEE-CEC 2015 optimization test bed on bound constrained single-objective computationally expensive numerical optimization. Remarkably, MVMO proves effective at solving different complex problems within a reduced number of allowed function evaluations.