Zhang, Jincheng · Zenodo (CERN European Organization for Nuclear Research) · 2026
In recent years, nature-inspired optimization algorithms have demonstrated powerful performance in continuous and combinatorial optimization problems. Traditional optimization algorithms often suffer from problems such as getting trapped in local optima, slow convergence speed, or insufficient search accuracy when dealing with high-dimensional, multi-modal complex functions. This paper proposes a novel swarm intelligence optimization algorithm—Adaptive Sand Rat Optimization (ASRO)—inspired by the foraging behavior, social cooperation, and risk-avoidance jumping mechanisms of gerbils in their natural environment. The algorithm achieves an organic combination of global search and local exploitation by introducing strategies such as dynamic environment perception, swarm cooperation and information sharing, nonlinear jumping, multi-level search, and adaptive convergence control. This paper presents the mathematical modeling, detailed process, and convergence control formula of the algorithm, and verifies it on several typical optimization test functions. Experimental results show that ASRO outperforms traditional swarm intelligence algorithms in terms of global search capability, local search accuracy, and convergence stability.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
Swarm intelligence optimization algorithms solve complex optimization problems by simulating the collective behavior of organisms in nature. This paper proposes a novel swarm intelligence algorithm—the Silverfish Optimization Algorithm (SFOA)—inspired by the rapid swimming, environmental perception, and cooperative behavior of silverfish in humid environments. The algorithm achieves an efficient balance between global search and local development by introducing global random exploration, neighbor cooperation, local fine-tuning, environmental perception mechanisms, and swarm memory mechanisms. This paper details the algorithm modeling, mathematical formula description, iterative process, characteristic analysis, and application potential, and theoretically analyzes the algorithm's convergence and swarm intelligence characteristics. Results show that SFOA possesses strong global search capabilities, fine-tuning local optimization capabilities, dynamic adaptive characteristics, and good swarm cooperation effects, making it suitable for various continuous and combinatorial optimization problems.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
omplex optimization problems by simulating the collective behavior of organisms in nature. This paper proposes a novel swarm intelligence optimization algorithm—the Nematomorpha Optimization Algorithm (NOA). The algorithm draws inspiration from the behavior of horsehair worms in nature, which involve crawling, contracting, extending, and sensing the environment to find water sources. It constructs a multi-stage crawling mechanism, an energy-information coupling mechanism, an environmentally sensitive perception mechanism, and a hierarchical group cooperation mechanism to achieve an effective balance between global search and local exploitation. In the algorithm design, mathematical formulas are introduced to accurately model the mechanisms of individual crawling, step size adjustment, energy updating, and group cooperation. The characteristics and potential application scenarios of the algorithm are analyzed. The results show that NOA has strong global search and local exploitation capabilities in high-dimensional continuous optimization problems, providing a new approach to swarm intelligence optimization algorithms.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
omplex optimization problems by simulating the collective behavior of organisms in nature. This paper proposes a novel swarm intelligence optimization algorithm—the Nematomorpha Optimization Algorithm (NOA). The algorithm draws inspiration from the behavior of horsehair worms in nature, which involve crawling, contracting, extending, and sensing the environment to find water sources. It constructs a multi-stage crawling mechanism, an energy-information coupling mechanism, an environmentally sensitive perception mechanism, and a hierarchical group cooperation mechanism to achieve an effective balance between global search and local exploitation. In the algorithm design, mathematical formulas are introduced to accurately model the mechanisms of individual crawling, step size adjustment, energy updating, and group cooperation. The characteristics and potential application scenarios of the algorithm are analyzed. The results show that NOA has strong global search and local exploitation capabilities in high-dimensional continuous optimization problems, providing a new approach to swarm intelligence optimization algorithms.
Cited by 0Open access
Zhang, Jincheng · Zenodo (CERN European Organization for Nuclear Research) · 2026
Swarm intelligence optimization algorithms have demonstrated excellent performance in solving complex optimization problems. Their core idea is to simulate the collective behavior of organisms in nature to achieve search and learning. This paper proposes a novel swarm intelligence optimization algorithm based on the foraging, migration, predator avoidance, and social behavior of snapper (SOA). This algorithm achieves dynamic search for individuals through four types of behaviors: local foraging, global migration, defense and avoidance, and social learning. It also introduces an energy adaptation mechanism, a social weight mechanism, and a historical trajectory reference strategy to improve the algorithm's global search capability and local exploration accuracy. This paper provides a detailed modeling and mathematical formulation of the algorithm, analyzing its convergence, global exploration capability, local refinement, and swarm intelligence characteristics. This algorithm is applicable to continuous optimization, multimodal functions, combinatorial optimization, and multi-objective optimization problems, and has broad application prospects.
Cited by 0Open access
Zhang, Jincheng · Zenodo (CERN European Organization for Nuclear Research) · 2026
Swarm optimization algorithms have demonstrated significant performance in continuous optimization, multimodal function solving, and complex constraint problems in recent years, including Particle Swarm Optimization (PSO), Grey Wolf Optimization (GWO), and Ant Colony Optimization (ACO). This paper proposes a novel swarm optimization algorithm—the Macaw Optimization Algorithm (MOA)—inspired by the foraging behavior, social learning, and environmental alertness of macaw colonies. MOA achieves an effective balance between global search and local fine-grained search by introducing adaptive exploration-development weights, a multi-leader parrot mechanism, dynamic social learning, an alertness mechanism, and an adaptive perturbation refinement strategy. This improves search efficiency and effectively avoids premature convergence. This paper provides a detailed mathematical model of the algorithm's principles and conducts experimental verification on a classic optimization test function. The results show that MOA outperforms traditional algorithms in both convergence speed and global optimum finding capability.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
To address the common problems of search direction degradation and lack of structured explanation of convergence behavior in existing swarm intelligence optimization algorithms for non-smooth objective functions, high-dimensional complex search spaces, and black-box optimization scenarios, this paper proposes a novel swarm intelligence optimization method—the Hedgehog Optimization Algorithm. Inspired by the defensive, probing, and migrating behaviors of hedgehogs, this algorithm goes beyond simple behavioral analogies, introducing mechanisms such as directional distribution functions, symbolic response operators, and cognitive memory fields to elevate swarm intelligence search into a structured optimization process that can be characterized in a function space. The algorithm does not rely on the gradient information of the objective function, achieving efficient solutions to non-smooth and non-convex problems through the adaptive evolution of the directional distribution. Furthermore, this paper formulates the Hedgehog Optimization Algorithm as a joint optimization problem of position variables and directional distribution, and theoretically analyzes its convergence behavior from the perspective of minimizing free energy. Experiments and theoretical analysis show that this method has good stability and generalization potential in complex optimization tasks.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
To address the common problems of search direction degradation and lack of structured explanation of convergence behavior in existing swarm intelligence optimization algorithms for non-smooth objective functions, high-dimensional complex search spaces, and black-box optimization scenarios, this paper proposes a novel swarm intelligence optimization method—the Hedgehog Optimization Algorithm. Inspired by the defensive, probing, and migrating behaviors of hedgehogs, this algorithm goes beyond simple behavioral analogies, introducing mechanisms such as directional distribution functions, symbolic response operators, and cognitive memory fields to elevate swarm intelligence search into a structured optimization process that can be characterized in a function space. The algorithm does not rely on the gradient information of the objective function, achieving efficient solutions to non-smooth and non-convex problems through the adaptive evolution of the directional distribution. Furthermore, this paper formulates the Hedgehog Optimization Algorithm as a joint optimization problem of position variables and directional distribution, and theoretically analyzes its convergence behavior from the perspective of minimizing free energy. Experiments and theoretical analysis show that this method has good stability and generalization potential in complex optimization tasks.
Cited by 0Open access
Jincheng Zhang · 2025
Cited by 0Open access
Zhang, Jincheng · Zenodo (CERN European Organization for Nuclear Research) · 2025
This paper proposes a novel nature-inspired optimization algorithm, the Toad Optimization Algorithm (TOA). Its design draws inspiration from the ecological behavior of toads, including foraging, Duchenne gland venom diffusion, defense mechanisms, group aggregation and dispersion, and humidity-sensitive migration. By simulating the behavioral characteristics of toads in their natural environment, the algorithm integrates local search, global migration, potential field repulsion, and adaptive group aggregation and dispersion mechanisms to solve continuous optimization problems. This paper provides a detailed explanation of the algorithm's mechanisms, mathematical formula derivation, population dynamics analysis, and theoretical convergence characteristics. It proposes three core mechanisms: a venom potential field mechanism, a humidity-adaptive migration mechanism, and a group dynamic aggregation and dispersion mechanism. This mechanism, combined with a dynamic weight update strategy, achieves a balance between global and local search. The algorithm's design fully reflects the behavioral characteristics of toads and provides new design ideas for heuristic optimization algorithms.
Cited by 0Open access
Jincheng Zhang · 2025
Cited by 0Open access
Hale, David Kivilcim · 2001
Cited by 0Open access
Jincheng Zhang · 2025
Cited by 0Open access
Jan H. Jensen (111108), Martin Willemoës (559349), Jakob R. Winther (559350) et al. · Figshare · 2015
<p>Example of energy convergence during the various macro cycles of the optimization algorithm for WT-PR in complex with TF-PR peptide. Each point along the graph corresponds to the energy (computed with the RosettaDock energy function) of the lowest in energy decoy out of 500 produced during each of the <i>K</i> steps.</p>
Cited by 0Open access
Карий, О., Окіпний, С., Karyy, O. et al. · ELARTU (Ternopil National Technical University) · 2024
Cited by 0
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
To address the problems of premature convergence, over-reliance on instantaneous optimal solutions, and static population structure inherent in existing swarm intelligence optimization algorithms for complex, multimodal, and non-stationary optimization problems, this paper proposes a novel swarm intelligence optimization method—the Grouse Optimization Algorithm (GOA). Inspired by the grouse's ground foraging, sudden take-off, population splitting and regrouping, and risk-avoidance behaviors exhibited in high-pressure ecosystems, GOA systematically reconstructs the traditional optimization paradigm from three levels: search target modeling, population structure organization, and individual behavioral decision-making. GOA no longer treats the global optimum as the sole attraction point but introduces a habitat center modeling method based on trajectory stability. It achieves dynamic adaptive control of the search process through an evolutionary trajectory credibility mechanism, a split-regrouping search structure, an anti-habitat disturbance strategy, and a behavioral state evolution graph. Theoretically, this algorithm constructs a stochastic behavioral evolution optimization framework oriented towards long-term adaptability, providing a new research approach for the design of swarm intelligence optimization algorithms.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
To address the common problems of existing metaheuristic optimization algorithms in complex continuous optimization problems, such as insufficient global search capability in the early stage, susceptibility to local optima in the middle stage, and low accuracy in the later stage, this paper proposes a novel swarm intelligence optimization method—the Adaptive Multi-Strategy Caracal Optimization Algorithm. This algorithm uses the hunting behavior of the caracal in nature as a biological prototype, mathematically abstracting its behaviors such as stealthy observation, visual locking, leaping hunting, and failure avoidance, and constructing an optimization framework that integrates multi-elite guidance, adaptive strategy selection, failure memory learning, and energy annealing mechanisms. During the search process, the algorithm no longer relies on a single optimal individual, but introduces a visual locking field to comprehensively model multiple elite solutions, thereby enhancing its overall perception of the complex search space. Simultaneously, through a multimodal leaping strategy pool and a self-evolving exploration-exploitation mechanism, the search strategy is dynamically evolved. Theoretical analysis shows that this algorithm, while maintaining strong global exploration capability, significantly improves the later convergence stability and search accuracy, providing a new and effective approach for solving high-dimensional, multimodal, and nonlinear optimization problems.
Cited by 0Open access
Dadang Setiawan, Suyanto Suyanto, Bayu Erfianto et al. · SSRN Electronic Journal · 2023
Cited by 1Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
This paper proposes a swarm intelligence optimization algorithm inspired by the foraging behavior of toucans, called the Toucan Optimization Algorithm (TOA). This algorithm simulates the toucans' global exploration, local exploitation, collaborative sharing, and environmental adaptation capabilities during foraging in their natural environment, achieving efficient solutions to complex optimization problems. TOA utilizes adaptive jump step size, multi-scale beak accuracy, collaborative weights, and a history memory mechanism, combined with Lévy flight and local scanning strategies, to balance global search and local exploitation, maintaining stable convergence in continuous optimization, combinatorial optimization, and multi-objective optimization problems. This paper details the algorithm modeling, iterative formulas, and mathematical mechanisms of TOA, and analyzes its algorithmic characteristics, providing a theoretical basis for its application in engineering optimization, intelligent control, and data mining.
Cited by 0Open access
Jincheng Zhang · Zenodo (CERN European Organization for Nuclear Research) · 2026
To address the problem that traditional swarm intelligence optimization algorithms generally rely on optimal individual guidance, visual assumptions, or explicit velocity models, a novel swarm intelligence optimization method—the hagfish optimization algorithm—is proposed. This algorithm is based on the non-visual survival behavior of deep-sea hagfish as a biological inspiration, constructing a novel optimization model from four aspects: information perception, memory mechanisms, group structure, and search dynamics. The hagfish optimization algorithm does not rely on optimal solution attraction, but instead achieves non-target-driven search through fitness change perception. Simultaneously, it introduces a mucus field memory mechanism in continuous space to achieve environment-level shared memory. At the group level, entanglement, breakage, and recombination mechanisms are designed to enhance the ability to escape local optima. At the dynamic level, a mucus-damped search model is constructed to suppress search oscillations. This algorithm significantly differs from existing classical swarm intelligence algorithms in its mechanism design, providing a new modeling approach for complex continuous optimization problems.
Cited by 0Open access