This analysis uncovers a major operational shift where advanced architectures actively serve as catalysts to break the historical dependency on the traditional EL pathway, successfully introducing novel, non-traditional active-learning mechanics such as autonomous LLM-guided strategy selectors, discrete classifier-based boundary tracks, and generative exploratory loops.
Software model optimization is a process that generates architecture alternatives aimed at improving quantifiable non-functional properties of software systems, such as performance and reliability. Multi-objective evolutionary algorithms are commonly used to explore the search space and help designers identify trade-offs among competing non-functional properties (e.g., through a Pareto front). However, such algorithms face efficiency challenges in complex software models and large design spaces, since evaluating the fitness (i.e., the quality) of each architecture requires analysis tools that become computationally expensive when repeatedly invoked during the search process. In this paper, we explore the construction of surrogate models based on regression techniques to approximate the outputs of these analysis tools at significantly lower computational cost, while maintaining reasonable output accuracy. Our experimental results suggest that surrogate models provide savings of up to $30\%$ in computational time and maintain the Pareto front quality provided by evolutionary algorithms. Also, we observed some differences in the architectural models produced by our approach. Overall, surrogate models constitute a promising approach for scaling multi-objective architecture optimization to larger spaces and complex architectural models.
The rapid growth of high-dimensional and heterogeneous datasets has intensified the demand for
scalable and robust optimization approaches in data analytics. Traditional deterministic
algorithms frequently suffer from premature convergence, initialization sensitivity, and reduced
performance in nonlinear or noisy environments. Nature-inspired algorithms (NIAs), including
evolutionary, swarm-based, and physics-inspired approaches, provide stochastic population
based mechanisms capable of navigating complex search spaces effectively. This study develops a
unified cross-domain benchmarking framework for evaluating NIAs across feature selection,
clustering, predictive modeling, and image segmentation tasks. A structured taxonomy is
introduced, and performance is analyzed along five dimensions: predictive accuracy, convergence
efficiency, computational complexity, robustness, and hybridization trade-offs. Mathematical
formulations of representative algorithms (Genetic Algorithms, Particle Swarm Optimization, Ant
Colony Optimization, Simulated Annealing) are incorporated to strengthen analytical rigor.
Comparative synthesis reveals that algorithm suitability is task-dependent: evolutionary
algorithms excel in feature selection, swarm-based approaches converge efficiently in clustering,
and hybrid models yield superior predictive performance at increased computational cost. The
proposed framework establishes standardized evaluation criteria to enhance reproducibility and
cross-domain comparability. The findings provide theoretical insight and practical guidance for
selecting nature-inspired optimization strategies in complex data-driven environments.
Unknown authors· International Journal of Com...· 0 citations
Sparse large-scale multi-objective optimization problems (sparse LSMOPs) are highly challenging due to the curse of dimensionality and the inherent sparsity of Pareto optimal solutions. Although co-evolutionary algorithms demonstrate significant potential, existing methods are frequently constrained by static resource allocation and undifferentiated variable grouping strategies. These limitations induce a severe mismatch between computational budgets and evolutionary states, making it difficult to effectively eliminate redundant noise variables while activating critical ones. To bridge these gaps, this paper proposes a heterogeneous population co-evolutionary algorithm (HPCEA) tailored for sparse LSMOPs. First, an adaptive population division strategy is introduced to dynamically adjust the scales of convergence and diversity sub-populations based on real-time evolutionary states, facilitating the on-demand allocation of computational resources. Second, a heterogeneous variable grouping strategy is designed: a correlation grouping approach captures structural correlation patterns in sparse space within the convergence sub-population, whereas a structural grouping approach unearths potential sparse patterns for the diversity sub-population. Finally, guided by these distinct architectures and sparse importance scores, heterogeneous genetic operators precisely manipulate variables to balance rapid convergence and broad exploration. Extensive experiments against six state-of-the-art algorithms across eight benchmark suites and three real-world scenarios demonstrate HPCEA’s superiority.
Expensive multiobjective optimization requires a diverse Pareto approximation under a severely limited evaluation budget. Existing surrogate-assisted algorithms often treat decision variables homogeneously and rely on either global or local models, causing inaccurate regional prediction or premature search concentration. This paper proposes HDS-LGSC, a hierarchical decision-space modeling and local–global surrogate coevolutionary algorithm. A lightweight graph-attention encoder learns variable interactions and partitions variables into global-dominant, locally coupled, and weakly related groups. A global radial-basis-function ensemble and local Gaussian processes guide cooperating populations, while an adaptive infill criterion combines hypervolume improvement, uncertainty, and decision-space novelty. Experiments on four benchmarks and a building energy-efficiency case demonstrate lower IGD, higher hypervolume, improved scalability, and robust prediction. Ablation results verify the complementary contributions of hierarchical modeling, dual surrogates, and composite infill selection.
Wenchao Pan, Junhan Li· 2026 8th International Confe...· 0 citations
BOCoDe is introduced, an open-source, PyTorch-native benchmark comprising 307 BBO problems, including 159 engineering design tasks and widely used synthetic and HPO benchmarks that establish a reproducible and extensible foundation for developing and evaluating BO methods that better reflect the demands of engineering design.
Rosen Yu, Christophe Hatterer, A. Narayanan et al.· 0 citations
Results show that integrating local search significantly enhances performance, while a principled method for setting hybrid parameters ensures robustness and reproducibility, highlighting the potential of combining mathematical programming techniques with evolutionary algorithms for high-dimensional many-objective optimization problems.
Regina C. L. C. de Sousa, Dênis E. C. Vargas, Elizabeth F. Wanner et al.· Journal of Heuristics· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
