Stochastic Adaptive Optimization: A New Metaheuristic Augmented with Adaptive Capability

Exploitation and exploration are two important aspects in metaheuristic as a stochastic optimization, so that many studies that developed a new metaheuristic focus on both aspects.On the other hand, adaptability is not less important than both aspects as the optimization progress may fall into stagnation.Unfortunately, this third aspect is less considered so that only few recent metaheuristics that are augmented with adaptability.This paper builds a new stochastic technique called as stochastic adaptive optimization (SAO).SAO employs multiple options for searches that are picked stochastically.Meanwhile, SAO is augmented with adaptability during its second walk out of two walks.The action in the second walk depends on the improvement status after the first walk.Exploration is picked if the stagnation occurs while exploitation is kept performed if improvement takes place.SAO is then challenged to answer 23 standard functions as unconstrained problems, 4 engineering design problems, and economic load dispatch (ELD) problem in Indonesia as constrained problems.Through this appraisal, it is benchmarked with five new techniques: carpet weaver optimization (CWO), orangutan optimization algorithm (OOA), tailor optimization algorithm (TOA), crayfish optimization algorithm (COA), and hiking optimization (HO).The result conveys that SAO outmatches CWO, OOA, TOA, COA, and HO in 16, 13, 13, 18, and 23 functions while it is competitive in handling the engineering design and ELD problems.