Scholarlink

<p>Aligning objectives and key results (OKRs) across heterogeneous collaboration networks—spanning corporate, academic, government, and research domains—remains a significant challenge due to intricate interdependencies and semantic relationships among diverse entities. Conventional alignment approaches frequently fail to accurately model these multi-relational structures, resulting in inefficient decision-making and misaligned strategies.</p><p> To address this issue, this study introduces the Improved Graph Random Multi-Relational Sandcat Swarm Coupled Attention Network (Imp-GRMR-2SCAN)—a hybrid framework designed to enhance precision and adaptability in OKR alignment. Imp-GRMR-2SCAN combines a random-coupled neural network (RCNN) that enhances feature robustness by random coupling with a multi-relational graph attention network (MRGAN) for discovering semantic dependencies between relational layers. The improved version of the model is optimized with the Improved Sandcat Swarm Optimization Algorithm (ISCSOA). It is a bio-inspired approach that replicates the adaptive hunting strategy of sand cats for finding the optimal neural parameters in an efficient manner. The suggested method leverages Correlation Coefficients and Min–Max Normalization for pre-processing data, and a critic-guided decision transformer (CGDT) for effective feature extraction. Experiments on the Driven Alignment OKR Dataset (500 records, 10 features) achieve accuracy at 99.9%, outperforming state-of-the-art models in terms of scalability as well as optimization efficiency. Through the integration of multi-relational learning, semantic attention, and swarm-based optimization, the Imp-GRMR-2SCAN solution provides an interpretable and resilient solution to the long-standing OKR misalignment issue in dynamic, data-intensive collaborative systems. The integration improves between intelligent network modeling and strategic goal alignment and plots a new path for data-driven organizational decision-making systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Improved graph random multi-relational sandcat swarm coupled attention network-driven alignment of objectives and key results through heterogeneous collaboration and semantic networks

  • 期刊论文

  • 发表日期：

    2025年12月9日

查看全文
  • Wenjian Xu,
  • Zulong Chen,
  • Ying Zhou,
  • Moufeng Zhang,
  • Yu Li,
  • Zhijie Jia,
  • Lei Zhang

摘要

Aligning objectives and key results (OKRs) across heterogeneous collaboration networks—spanning corporate, academic, government, and research domains—remains a significant challenge due to intricate interdependencies and semantic relationships among diverse entities. Conventional alignment approaches frequently fail to accurately model these multi-relational structures, resulting in inefficient decision-making and misaligned strategies.

To address this issue, this study introduces the Improved Graph Random Multi-Relational Sandcat Swarm Coupled Attention Network (Imp-GRMR-2SCAN)—a hybrid framework designed to enhance precision and adaptability in OKR alignment. Imp-GRMR-2SCAN combines a random-coupled neural network (RCNN) that enhances feature robustness by random coupling with a multi-relational graph attention network (MRGAN) for discovering semantic dependencies between relational layers. The improved version of the model is optimized with the Improved Sandcat Swarm Optimization Algorithm (ISCSOA). It is a bio-inspired approach that replicates the adaptive hunting strategy of sand cats for finding the optimal neural parameters in an efficient manner. The suggested method leverages Correlation Coefficients and Min–Max Normalization for pre-processing data, and a critic-guided decision transformer (CGDT) for effective feature extraction. Experiments on the Driven Alignment OKR Dataset (500 records, 10 features) achieve accuracy at 99.9%, outperforming state-of-the-art models in terms of scalability as well as optimization efficiency. Through the integration of multi-relational learning, semantic attention, and swarm-based optimization, the Imp-GRMR-2SCAN solution provides an interpretable and resilient solution to the long-standing OKR misalignment issue in dynamic, data-intensive collaborative systems. The integration improves between intelligent network modeling and strategic goal alignment and plots a new path for data-driven organizational decision-making systems.