metaheuristic hyperparameter tuning

In Python search-based unit test generation, the DynaMOSA and MIO algorithms suffer from inefficient hyperparameter configurations; their default settings often yield suboptimal code coverage, while conventional grid search incurs excessive computational overhead. Method: This paper introduces differential evolution (DE) into the Pynguin framework for automated multi-objective hyperparameter optimization of search-based test generation algorithms. We design a fitness function and encoding scheme tailored to testing objectives and conduct end-to-end tuning experiments on standard benchmarks. Contribution/Results: DE-optimized DynaMOSA achieves significant improvements in branch and line coverage (average +8.2%) over baseline configurations. Moreover, it converges 3.7× faster than grid search while reducing tuning overhead by approximately 65%. This work establishes an efficient, reproducible hyperparameter optimization paradigm for search-based test generation.

This study systematically evaluates the robustness and structural invariance of hybrid population-based metaheuristics under objective-space transformations—including translation, scaling, rotation, and composite deformations. We propose a lightweight, plug-and-play generic hybrid framework that seamlessly integrates 19 state-of-the-art algorithms and conduct comprehensive multi-dimensional comparative experiments on the CEC-2017 benchmark suite. Leveraging Wilcoxon and Friedman nonparametric tests, Bayesian dominance analysis, and convergence trajectory profiling, we find that DE-based hybrids (e.g., hIMODE, hSHADE) significantly outperform PSO- and bio-inspired methods in stability and invariance—particularly in high-dimensional, non-separable, and rotation/composite-deformation scenarios, where they maintain high accuracy and strong robustness. This work is the first to empirically uncover the intrinsic structural advantages of DE-family algorithms under geometric distortions, providing both theoretical foundations and practical design principles for optimization algorithms targeting complex, dynamically transforming environments.

This study addresses the lack of theoretical foundation for parameter selection in the bat algorithm, which has traditionally relied on empirical tuning. For the first time, it integrates dynamical systems theory with population variance evolution analysis to construct a theoretical framework characterizing the influence of key parameters. Within this framework, analytically derived effective ranges for critical parameters are established. Numerical experiments confirm that the theoretical predictions align closely with the observed convergence behavior in practice. The work further uncovers the intrinsic mechanisms governing the trade-off between exploration and exploitation and the algorithm’s convergence properties, thereby providing the first systematic theoretical guidance for parameter configuration in the bat algorithm.

Manual design of metaheuristic algorithms is time-consuming, inefficient, and structurally inflexible; existing automated approaches are constrained by fixed algorithmic templates and linear representations. Method: This paper proposes the first general-purpose automated design framework applicable to the entire family of metaheuristics. It introduces (1) a unified algorithmic prototype covering all metaheuristic variants; (2) a directed acyclic graph (DAG)-based representation enabling structural evolution; and (3) a compact yet expressive graph embedding and differentiable architecture encoding scheme. Integrating graph representation learning with evolutionary search, the framework enables end-to-end generation of diverse, nonlinear algorithmic structures. Results: Extensive evaluation on numerical optimization benchmarks and real-world tasks demonstrates that the automatically generated algorithms achieve superior efficiency, generalizability, and novelty compared to both human-designed and state-of-the-art automated methods.

To address the exploration-exploitation imbalance and inefficient utilization of historical evaluation data in evolutionary algorithms (EAs) for hyperparameter optimization (HPO), this paper proposes an enhanced genetic algorithm (GA) framework integrated with a lightweight linear surrogate model. The linear surrogate is seamlessly embedded into GA’s selection and mutation operators, enabling a population-driven hybrid search strategy and performance-feedback-driven adaptive model updating—without requiring gradient computation or expensive modeling overhead. This design dynamically balances global exploration and local exploitation. On standard HPO benchmarks, the method achieves an average performance improvement of 1.89% (range: −3.45% to +6.55%) over state-of-the-art approaches, with negligible increase in training cost. Its core contribution lies in the first structured integration of linear surrogates with genetic operations, achieving a favorable trade-off among efficiency, accuracy, and scalability.

This study addresses the challenge that high-performing metaheuristic algorithms for the Vehicle Routing Problem (VRP) often rely heavily on manual parameter tuning and domain-specific expertise. To overcome this limitation, the authors propose a Metacognitive Evolutionary Programming (MEP) framework, which introduces a large language model as a strategic discovery agent. Within a Reason–Act–Reflect loop, this agent actively diagnoses, hypothesizes, and refines core heuristic components of the Hybrid Genetic Search (HGS) algorithm. Unlike conventional approaches that employ passive feedback mechanisms, MEP enables explicit reasoning about the exploration–exploitation trade-off and facilitates heuristic innovation. Experimental results across multiple complex VRP variants demonstrate that MEP improves solution quality by up to 2.70% compared to the original HGS while reducing runtime by more than 45%.

Existing automated algorithm design frameworks (e.g., EoH, FunSearch) optimize only algorithmic structure while neglecting systematic prompt evolution, limiting LLM performance on complex NP-hard problems. Method: We propose the first LLM-driven co-evolutionary framework that jointly optimizes population-based metaheuristic algorithms and guiding prompts. Our approach employs a unified large language model (GPT-4o-mini/Qwen3-32B/GPT-5) to instantiate a co-evolutionary mechanism integrating population intelligence modeling with interpretable, structured prompt template evaluation. Contribution/Results: The framework breaks the static prompt dependency bottleneck and reduces reliance on high-end LLMs. It achieves significant improvements over state-of-the-art methods across multiple NP-hard problem classes. Ablation studies confirm the necessity of co-evolution, and cross-model evolutionary trajectory analysis reveals intrinsic coupling mechanisms between prompts and algorithmic components.

This work addresses the long-standing challenge that reinforcement learning-based hyper-heuristics (RLHH) often struggle to effectively select low-level heuristics on standard benchmark functions. Focusing on the LeadingOnes problem and combining two randomized local search operators—RLS₁ and RLS₂—the study provides the first rigorous theoretical proof, supported by empirical validation, that RLHH can achieve the optimal expected runtime attainable by these two operators (up to lower-order terms) under appropriate parameter settings. This result refutes prior skepticism regarding RLHH’s capacity to learn effective heuristic selection policies. Furthermore, experimental results demonstrate that, at practical problem scales, RLHH outperforms the generalized randomized gradient hyper-heuristic, which also enjoys optimal theoretical runtime guarantees.

This study addresses the Traveling Salesman Problem (TSP), a classic NP-hard combinatorial optimization challenge, by proposing a high-order relay hybrid solving framework. The approach innovatively integrates the swarm intelligence-based Dragonfly Algorithm (DA) with the memory-driven Tabu Search (TS): DA first performs global exploration to generate high-quality initial solutions, which are subsequently refined through local search via TS. Coupled with a systematic grid-search parameter tuning strategy, the proposed framework demonstrates superior performance on standard TSPLIB instances compared to standalone DA or TS, and outperforms classical metaheuristics such as Genetic Algorithms and Ant Colony Optimization. The results indicate significant improvements in both solution quality and robustness.

This work proposes a two-level deep reinforcement learning framework for large-scale Traveling Salesman Problems (TSP), wherein a recurrent Proximal Policy Optimization (PPO) agent dynamically controls both numerical and structural parameters of a genetic algorithm, enabling their decoupled analysis. The study provides the first empirical evidence that dynamic adjustment of structural parameters is crucial for avoiding premature convergence and escaping local optima, whereas numerical parameters serve only a fine-tuning role. Evaluated on large-scale TSP instances such as rl5915, the proposed method significantly outperforms static baselines, reducing the optimality gap by approximately 45%. These results offer a novel direction for automated algorithm design through adaptive parameter control in evolutionary computation.