Camel foraging optimization algorithm: a mechanism-driven metaheuristic with fractional memory for constrained optimization

Idriss Dagal

doi:10.1038/s41598-026-67080-1

Camel foraging optimization algorithm: a mechanism-driven metaheuristic with fractional memory for constrained optimization

2026-08-21

0

OA

AI

DOI：10.1038/s41598-026-67080-1

Original

Original request for help

Save

Abstract

En 中文

Global optimization problems often involve nonlinear, multimodal, non-separable, and constrained search landscapes that challenge population-based metaheuristics under limited function-evaluation budgets. This paper proposes the camel foraging optimization algorithm (CFEOA), a mechanism-driven metaheuristic that combines endurance-regulated state control with bounded Grünwald–Letnikov fractional-order displacement memory. In CFEOA, endurance acts as an agent-specific variable regulating the transition between exploration and local refinement, while the bounded fractional-memory term introduces directional persistence from recent accepted displacements. The camel-foraging analogy is used only as an organizing metaphor; the search process is defined through explicit mathematical operators, boundary repair, acceptance rules, and function-evaluation accounting. CFEOA is evaluated using 30 independent runs, fixed random seeds, equal function-evaluation budgets, disabled early stopping in the main benchmark comparisons, full parameter disclosure, and public code availability. The main validation uses the CEC 2022 suite with classical swarm baselines and adaptive DE-family comparators, including NL-SHADE-LBC as an additional recent SOTA reference. Across 18 function–dimension blocks, NL-SHADE-LBC achieves the most favorable aggregate rank, whereas CFEOA retains statistically significant advantages over PSO, GWO, and WOA and shows no significant difference from DE, JADE, SHADE, L-SHADE, and TLBO after Holm correction. Component-wise ablation, parameter sensitivity, convergence, diversity, CPU-time, scalability, and constrained engineering analyses indicate that CFEOA provides a transparent, reproducible, and reliability-oriented search framework with landscape-dependent strengths rather than universal SOTA dominance. The results position CFEOA as a controlled-search alternative for selected complex and constrained optimization settings where repeatability, feasibility, and auditable mechanism design are important.

Camel foraging optimization algorithm: a mechanism-driven metaheuristic with fractional memory for constrained optimization

Abstract

Journal

Organization