PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
May 26, 2026Journal of Computing in Civil Engineering0 citations

DfMA-Oriented Rebar Design Optimization for Reinforced Concrete Slabs Using a Hybrid Mechanism Integrating Graph Neural Networks and Metaheuristic Algorithms

View Full Paper
MLMingkai LiFLFangzhou LinTWTao Wang

Key Points

  • The aim is to optimize rebar design in reinforced concrete slabs for better performance and efficiency using a hybrid optimization method.
  • Developed a hybrid optimization framework integrating graph neural networks and metaheuristic algorithms.
  • Used metaheuristic algorithms to generate optimal rebar layouts for various slab cases.
  • Trained the graph neural network to predict rebar configurations based on new slab geometries followed by refinement and compliance checking.
  • Achieved 40%–70% time savings in computational efficiency compared to traditional methods.
  • Demonstrated significant improvements in cost-effectiveness and constructability of rebar designs.
  • Validated through multiple case studies showcasing enhanced design quality.

Abstract

Reinforced concrete (RC) slabs are fundamental structural components in buildings and infrastructure, providing load-bearing capacity and overall stability. Optimizing rebar design for RC slabs is crucial for balancing structural performance, material efficiency, and constructability. Traditional design methods are often oversimplified and overlook the complex spatial distribution of rebar, leading to material waste. Existing optimization methods usually focus on minimizing material consumption, resulting in challenges in practical implements. To address these limitations, this paper proposes a hybrid optimization framework that integrates (GNNs) and metaheuristic algorithms (MAs) under design for manufacture and assembly principles. In the proposed approach, the MA module first generates optimal rebar layouts for multiple slab cases, which are used to construct and train the GNN model. The trained GNN then predicts an initial rebar configuration based on new slab geometries, and the MA subsequently performs localized refinement and code compliance checking. This hybrid approach enhances computational efficiency (40%–70% time saving) and design quality by systematically optimizing rebar amount and distribution. The proposed method is validated through a series of case studies, demonstrating significant improvements in cost-effectiveness, constructability, and computational efficiency compared with conventional approaches.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Li et al. (2026) studied this question.

synapsesocial.com/papers/6a153b00b5d9c58d83e8d446https://doi.org/10.1061/jccee5.cpeng-7477
Ask AI
Helpful
Bookmark
Share
View Full Paper