Scholarlink

<p>During sugarcane extraction operations, key process parameters are often adjusted manually by on-site operators based on experience, resulting in significant randomness and energy waste. To overcome these limitations, this study develops a hybrid-model-based optimization framework to improve the efficiency of the extraction process. First, a material and energy flow interaction model is established by incorporating collaboration theory to simulate the extraction process and derive ordered state parameters of the flow subsystems. These parameters are then used as inputs for multiple data-driven models, including long short-term memory (LSTM), extreme gradient boosting (XGBoost), and deep kernel extreme learning machine (DKELM), which jointly predict extraction rate and energy consumption. To enhance prediction accuracy and reduce local errors, a Bayesian model averaging (BMA) approach integrates the outputs of these submodels, forming an accurate and robust flow model. Based on this integrated model, a multi-objective optimization mechanism is further developed to simultaneously maximize extraction rate and minimize specific energy consumption, thereby improving overall system performance. A coati optimization algorithm guided by flow coordination is applied to fine-tune operational parameters under varying working conditions. Experimental results verify the effectiveness of the proposed approach, showing a 4.92% reduction in energy consumption per ton of sugarcane, a 0.09% increase in extraction rate, and a 67.92% improvement in collaborative performance.</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

Bi-objective optimization of sugarcane milling process based on weighted ensemble learning

  • 期刊论文

  • 发表日期：

    2026年3月21日

查看全文
  • Jie Chen,
  • Yiwen Hu

摘要

During sugarcane extraction operations, key process parameters are often adjusted manually by on-site operators based on experience, resulting in significant randomness and energy waste. To overcome these limitations, this study develops a hybrid-model-based optimization framework to improve the efficiency of the extraction process. First, a material and energy flow interaction model is established by incorporating collaboration theory to simulate the extraction process and derive ordered state parameters of the flow subsystems. These parameters are then used as inputs for multiple data-driven models, including long short-term memory (LSTM), extreme gradient boosting (XGBoost), and deep kernel extreme learning machine (DKELM), which jointly predict extraction rate and energy consumption. To enhance prediction accuracy and reduce local errors, a Bayesian model averaging (BMA) approach integrates the outputs of these submodels, forming an accurate and robust flow model. Based on this integrated model, a multi-objective optimization mechanism is further developed to simultaneously maximize extraction rate and minimize specific energy consumption, thereby improving overall system performance. A coati optimization algorithm guided by flow coordination is applied to fine-tune operational parameters under varying working conditions. Experimental results verify the effectiveness of the proposed approach, showing a 4.92% reduction in energy consumption per ton of sugarcane, a 0.09% increase in extraction rate, and a 67.92% improvement in collaborative performance.