Scholarlink

<p>In the Industry 5.0 era, intelligent negotiation requires systems that combine computational precision with human-centric insights, including sentiment, trust, and strategic adaptability. This study introduces an intelligent negotiation framework that integrates deep reinforcement learning, natural language processing, and MCDM to support adaptive, emotionally aware, and transparent decision-making. A deep reinforcement learning model using the proximal policy optimization algorithm learns optimal negotiation strategies under uncertainty, while advanced transformer-based models such as BERT for semantic understanding, RoBERTa for emotion detection, and SBERT for conversational coherence enable nuanced interpretation of negotiation dynamics. A hybrid AHP-TOPSIS method ranks proposals based on both technical and human-centric criteria, such as cost, trust in AI, and decision stability. A dynamic feedback loop further refines learning based on historical outcomes. Experimental simulations show that the integrated model improves negotiation success rates, enhances sentiment detection accuracy, reduces agreement costs, and increases the acceptance of optimal proposals. Sensitivity analysis confirms stable performance across diverse negotiation conditions. By bridging strategic learning, emotional intelligence, and multi-criteria optimization, the proposed framework offers a scalable, interpretable, and robust approach to automated negotiation in domains such as supply chains, e-commerce, and AI-enabled contract management. This work advances automated negotiation by addressing both algorithmic rigor and human-oriented requirements in complex decision environments.</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

An Intelligent Negotiation Framework for Industry 5.0: Integrating Deep Reinforcement Learning, Natural Language Processing, and MCDM

  • 期刊论文

  • 发表日期：

    2026年4月3日

查看全文
  • Aram Bahrini,
  • Amir Aghsami,
  • Behnam Malmir

摘要

In the Industry 5.0 era, intelligent negotiation requires systems that combine computational precision with human-centric insights, including sentiment, trust, and strategic adaptability. This study introduces an intelligent negotiation framework that integrates deep reinforcement learning, natural language processing, and MCDM to support adaptive, emotionally aware, and transparent decision-making. A deep reinforcement learning model using the proximal policy optimization algorithm learns optimal negotiation strategies under uncertainty, while advanced transformer-based models such as BERT for semantic understanding, RoBERTa for emotion detection, and SBERT for conversational coherence enable nuanced interpretation of negotiation dynamics. A hybrid AHP-TOPSIS method ranks proposals based on both technical and human-centric criteria, such as cost, trust in AI, and decision stability. A dynamic feedback loop further refines learning based on historical outcomes. Experimental simulations show that the integrated model improves negotiation success rates, enhances sentiment detection accuracy, reduces agreement costs, and increases the acceptance of optimal proposals. Sensitivity analysis confirms stable performance across diverse negotiation conditions. By bridging strategic learning, emotional intelligence, and multi-criteria optimization, the proposed framework offers a scalable, interpretable, and robust approach to automated negotiation in domains such as supply chains, e-commerce, and AI-enabled contract management. This work advances automated negotiation by addressing both algorithmic rigor and human-oriented requirements in complex decision environments.