Attention-Boundary Algorithm: A New Metaheuristic and Its Application in Handling Engineering Design Problem

This work introduces a new metaheuristic called attention-boundary algorithm (ABA).It offers a new approach in exploiting the boundaries of the space.The boundaries have 50 percent probability to be chosen as initial solution while the rest 50 percent probability is uniformly distributed within the space.Then, there is 50 percent probability that a solution is restarted at the boundaries if stagnation takes place.ABA employs three references for its directed searches which are the randomly solution within space, the middle of all better units plus the most optimal unit, and a randomly chosen unit within swarm.ABA is then implemented to find the optimal solution of 23 standard functions and four engineering design problems including pressure vessel, speed reducer, welded beam, and spring design problems.In this assessment, ABA is confronted with five metaheuristics including carpet weaver optimization (CWO), dollmaker optimization algorithm (DOA), hiking optimization (HO), total interaction optimization (TIA), and pufferfish optimization algorithm (POA).The result reveals that ABA is better than CWO, DOA, HO, TIA, and POA in 21,17,23,18,and 16 functions consecutively in managing 23 standard functions.Meanwhile, ABA is on the best in handling speed reducer, third best in handling the pressure vessel design problem, and on the fourth best in handling spring and welded beam design problems.