Probabilistic Ranking Grey Wolf Optimizer Improves Multi-Objective Collaborative Logistics Network Calculations

A research team has introduced a novel evolutionary algorithm designed to solve complex multi-objective optimization problems in spatio-temporal adaptive collaborative logistics networks. Investigated by researchers Yitong Liu, Tianyu Li, and Hong Zhang, the optimization challenges in these collaborative networks typically require balancing multiple competing targets simultaneously. The newly proposed method, designated as the Surrogate-assisted Evolutionary Multi-objective Grey Wolf Optimizer with Probabilistic Ranking Strategy, or SAE-MOGWO-PRS, seeks to resolve persistent computational bottlenecks that arise during traditional evolutionary search processes.

In conventional practice, most evolutionary multi-objective optimization algorithms fail to account for heterogeneity in evaluation times across distinct objectives. Standard surrogate-assisted optimization frameworks generally identify promising population positions across the broader solution space and then execute high-fidelity evaluations across every individual objective function. Although performing exhaustive high-fidelity evaluations on all objectives provides complete and error-free assessments of candidate solutions, the researchers emphasized that this conventional approach does not yield the most computationally efficient procedure when processing times vary.

To overcome these computational limitations, the SAE-MOGWO-PRS framework incorporates a probabilistic ranking strategy specifically tailored to manage heterogeneous objective evaluation demands. The algorithm is structured to determine selectively which candidate individuals and which specific objective functions within the population ought to undergo computationally demanding high-fidelity evaluations. By strategically targeting these evaluations rather than assessing all objectives uniformly, the method enhances the capacity of the Multi-objective Grey Wolf Optimizer to generate non-dominated solutions characterized by improved convergence behavior and broader population diversity.

The authors validated the operational efficacy of the SAE-MOGWO-PRS algorithm by benchmarking its performance against four state-of-the-art surrogate-assisted multi-objective optimization algorithms. This comparative evaluation explicitly included competing algorithms engineered to handle heterogeneous evaluation times. Across a suite of four benchmark test problems, comparative analysis demonstrated the superior performance of the probabilistic ranking strategy-based grey wolf optimizer relative to the baseline methods, confirming the overall effectiveness of the proposed surrogate-assisted methodology for multi-objective optimization.

Subject of Research: Multi-objective optimization algorithms for spatio-temporal adaptive collaborative logistics networks

Article Title: SAE-MOGWO-PRS: A probability ranking strategy-based grey wolf optimizer for spatio-temporal adaptive collaborative logistics network optimization

Article References: Liu, Y., Li, T., & Zhang, H. (2026). SAE-MOGWO-PRS: A probability ranking strategy-based grey wolf optimizer for spatio-temporal adaptive collaborative logistics network optimization. PLOS One, 21(11), e0362456. https://doi.org/10.1371/journal.pone.0362456

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0362456

Keywords: grey wolf optimizer, surrogate-assisted optimization, multi-objective optimization, probabilistic ranking strategy, collaborative logistics networks, spatio-temporal optimization, evolutionary algorithms, heterogeneous evaluation times, high-fidelity evaluations, non-dominated solutions, algorithm convergence, solution diversity, computational efficiency, population positions, test problem analysis, comparative algorithm benchmark, logistics network optimization