Velvet Worm Optimization: A Novel Bio-inspired Metaheuristic for Engineering Design Problems

This study introduces a novel bio-inspired metaheuristic algorithm, Velvet Worm Optimization (VWO), motivated by the unique behavioral repertoire of velvet worms (Onychophora).VWO translates the worm's natural strategies-slow sensory exploration, directional motion guided by antennae, cooperative interactions, and targeted slime-shooting predation-into mathematically defined operators that integrate global exploration and local exploitation.During the exploration phase, the algorithm emulates antennal probing and environmental scanning through adaptive random walks and population-based information sharing.This facilitates a comprehensive coverage of the search space, preserves diversity, and mitigates premature convergence.As iterations advance, agents progressively bias their movement toward more promising regions based on collective knowledge, achieving a smooth transition from stochastic exploration to guided search.In the exploitation phase, VWO models the precision of slime shooting as a localized intensification process.Each agent refines its search around the best-known positions, employing a probability-driven directional motion to enhance convergence toward near-optimal solutions.The performance of the algorithm is evaluated on four classical engineering design problems, namely tension/compression spring, welded beam, speed reducer, and pressure vessel, and it is compared against nine state-of-the-art metaheuristics.Experimental results indicate that VWO consistently delivers a superior or comparable solution quality, faster convergence, and reduced variability, attaining top ranks in three out of four problems, thereby demonstrating robustness, efficiency, and practical applicability to complex constrained optimization tasks.