Improved black-winged kite optimization algorithm with multi-strategy hybrid and its application

Key findings from this study

• The study found that integrating optimal point set initialization, adaptive attack weighting, alert behaviors, and Levy flight migration substantially improved the Black-winged kite algorithm's convergence characteristics.
• The authors report that Markov chain analysis provided theoretical validation of convergence properties for the improved algorithm.
• The researchers demonstrate that IMBKA outperformed five established optimization algorithms in standardized benchmark testing.

Overview

The study addresses convergence limitations in the Black-winged kite optimization algorithm (BKA) by developing an improved multi-strategy hybrid variant (IMBKA). The proposed algorithm integrates optimal point set initialization, adaptive weighting for attack behaviors, alert mechanisms, and Levy flight-enhanced migration to reduce premature convergence to local optima.

Methods and approach

The researchers enhanced BKA through four primary modifications. Initial population generation utilized an optimal point set model for improved diversity. Attack behavior incorporated adaptive weighting to balance exploration-exploitation tradeoffs. Alert behaviors were integrated to strengthen robustness. Levy flight strategy combined with migration behavior addressed local optima entrapment. Markov chain analysis established theoretical convergence properties. Comparative benchmarking evaluated IMBKA against five alternative algorithms using standard test functions. A Support Vector Machine parameter optimization task applied the algorithm to pantograph-catenary contact resistance prediction.

Results

IMBKA demonstrated superior performance relative to five comparative algorithms across test function evaluations. The benchmarking results showed improved convergence characteristics and solution quality across multiple optimization landscapes. Application to SVM parameter optimization for pantograph-catenary contact resistance prediction produced model predictions that validated the algorithm's practical utility in engineering contexts.

Implications

The multi-strategy integration approach establishes a methodology for enhancing population-based optimization algorithms through complementary mechanisms. Improved convergence properties extend applicability to complex engineering parameter optimization problems where local optima represent significant obstacles. The pantograph-catenary application demonstrates transferability to industrial predictive modeling scenarios requiring robust hyperparameter tuning.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.