Stochastic Boundary Algorithm: A New Metaheuristic with Adaptive Capability and Boundary Tracing

A novel optimization technique called stochastic boundary algorithm (SBA) is presented in this paper.This technique is metaphor-free and constructed based on swarm intelligence (SI) framework.It provides novel approach through adaptive capability, multiple traces, and exploitation of the boundary.Adaptive capability is conducted by performing exploitative trace when improvement occurs and explorative trace when stagnation takes place.SBA is a multiple trace technique where it selects the trace stochastically.The boundary trace is conducted as a part of exploitative trace despites the local trace with declining space representing the trace that moves from exploration to exploitation.The investigation is conducted by employing SBA to tackle a set of 23 functions as unconstrained case and a set of four engineering design problems as constrained case.Five techniques are chosen as comparators including spider-tailed horn viper optimization (STHVO), carpet weaver optimization (CWO), particle swarm optimization (PSO), red panda optimization (RPO), and crayfish optimization algorithm (COA).The result shows that SBA is dominant in tackling high dimension functions and competitive in tackling fixed dimension functions and engineering design problems.SBA is more prominent than STHVO, CWO, PSO, RPO, and COA in 12,14,23,11,and 21 functions.SBA is on the second position in tackling pressure vessel design problem and fourth position in tackling spring, welded beam, and speed reducer design problems.