Scholarlink

<p>Modern construction project planning increasingly requires the simultaneous optimization of multiple conflicting objectives, particularly time, cost, quality, and safety risk. This study presents an artificial intelligence-based multi-objective optimization framework for addressing the construction time–cost–quality–safety risk trade-off (TCQSRT) problem. A 37-activity project is considered as a case study, where each activity is associated with multiple execution modes characterized by different combinations of duration, cost, quality, and safety risk, resulting in a large-scale combinatorial optimization problem. To effectively solve this complex problem, an enhanced Aquila Optimizer (OBL–AO) integrated with opposition-based learning is proposed. The Aquila Optimizer, inspired by the hunting behavior of Aquila eagles, provides a balanced exploration and exploitation mechanism, while opposition-based learning improves population diversity and accelerates convergence by simultaneously evaluating candidate solutions and their opposite solutions. The proposed model simultaneously minimizes project duration, total cost, and safety risk while maximizing quality, generating a diverse set of Pareto-optimal solutions for decision-making. The performance of the proposed OBL–AO approach is evaluated using standard multi-objective metrics, including Spacing (Sp), Mean Ideal Distance (MID), and Hypervolume (HV). The results indicate that OBL–AO achieves a Sp value of 0.647, reflecting a competitive distribution of solutions across the Pareto front. In terms of convergence, it attains a MID value of 2.01, which is lower than NDSII-TLBO (2.21) and close to OBL-TLBO (1.97), demonstrating strong proximity to the ideal solution set. Most notably, OBL–AO achieves the highest HV value of 0.79 compared to OBL-TLBO (0.77) and NDSII-TLBO (0.74), indicating superior overall performance in terms of convergence and diversity balance. Overall, comparative analysis confirms that the proposed OBL–AO framework generates a more robust and well-distributed Pareto front than benchmark methods. These results demonstrate its effectiveness in producing high-quality trade-off solutions for the TCQSRT problem and highlight its potential as a reliable decision-support tool for improving sustainability, safety, and efficiency in modern construction project management.</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

Artificial intelligence for multi-objective optimization of time, cost, quality, and safety risks in construction projects

  • 期刊论文

  • 发表日期：

    2026年5月27日

查看全文
  • Vishnu Lakkamraju,
  • Gyana Ranjana Panigrahi,
  • Ramesh Chandra Sahoo,
  • Tushar Savale,
  • Tarun Madan Kanade,
  • Abdul Habib Zaray

摘要

Modern construction project planning increasingly requires the simultaneous optimization of multiple conflicting objectives, particularly time, cost, quality, and safety risk. This study presents an artificial intelligence-based multi-objective optimization framework for addressing the construction time–cost–quality–safety risk trade-off (TCQSRT) problem. A 37-activity project is considered as a case study, where each activity is associated with multiple execution modes characterized by different combinations of duration, cost, quality, and safety risk, resulting in a large-scale combinatorial optimization problem. To effectively solve this complex problem, an enhanced Aquila Optimizer (OBL–AO) integrated with opposition-based learning is proposed. The Aquila Optimizer, inspired by the hunting behavior of Aquila eagles, provides a balanced exploration and exploitation mechanism, while opposition-based learning improves population diversity and accelerates convergence by simultaneously evaluating candidate solutions and their opposite solutions. The proposed model simultaneously minimizes project duration, total cost, and safety risk while maximizing quality, generating a diverse set of Pareto-optimal solutions for decision-making. The performance of the proposed OBL–AO approach is evaluated using standard multi-objective metrics, including Spacing (Sp), Mean Ideal Distance (MID), and Hypervolume (HV). The results indicate that OBL–AO achieves a Sp value of 0.647, reflecting a competitive distribution of solutions across the Pareto front. In terms of convergence, it attains a MID value of 2.01, which is lower than NDSII-TLBO (2.21) and close to OBL-TLBO (1.97), demonstrating strong proximity to the ideal solution set. Most notably, OBL–AO achieves the highest HV value of 0.79 compared to OBL-TLBO (0.77) and NDSII-TLBO (0.74), indicating superior overall performance in terms of convergence and diversity balance. Overall, comparative analysis confirms that the proposed OBL–AO framework generates a more robust and well-distributed Pareto front than benchmark methods. These results demonstrate its effectiveness in producing high-quality trade-off solutions for the TCQSRT problem and highlight its potential as a reliable decision-support tool for improving sustainability, safety, and efficiency in modern construction project management.