Tapeworm foraging optimization algorithm: a novel approach to metaheuristic optimization of complex problems

Abstract This study presents an empirical investigation of the Tapeworm Foraging Optimization (TFO) algorithm, a nature-inspired metaheuristic search framework motivated by parasitic tapeworm attachment and reproductive segmentation. TFO combines a decaying Gaussian perturbation for exploratory diffusion with an inverted difference-vector displacement toward the current global best for local exploitation, governed by a probabilistic switching parameter. We evaluate TFO on classical unimodal and multimodal benchmarks (F1–F16) and four constrained mechanical engineering design problems: tension/compression spring, welded beam, 10-bar cantilever truss, and speed reducer design. Comparative evaluations against established metaheuristics - Grey Wolf Optimizer (GWO), Whale Optimization Algorithm (WOA), and Aquila Optimizer (AOA) reveal that GWO, WOA, and AOA converge faster and achieve objective values closer to theoretical minima on most benchmark functions, whereas TFO achieves comparable performance on F6 and F14 and competitive convergence on F11, F15, and F16. In constrained engineering optimization, TFO consistently identifies valid feasible configurations satisfying nonlinear stress, deflection, and geometric constraints, although it obtains higher objective costs than comparator methods on the spring (+ 7.14%), welded-beam (+ 8.57%), and 10-bar truss (+ 25.22%) problems, while securing an intermediate 3rd-place ranking on the speed reducer problem. Systematic parameter-sensitivity and ablation analyses across 1,980 independent runs indicate that balanced hybrid switching is essential for avoiding premature constraint entrapment, while the decay schedule exhibits robust stability. TFO’s structural simplicity, absence of velocity vectors or memory matrices, and low per-iteration overhead make it an accessible, transparent baseline for constrained engineering applications where computational simplicity and constraint feasibility are prioritized.