Development of an integrated intelligent BIM-based model for multi-objective optimization in engineering assembly...

Area of Science:

• Construction Management
• Artificial Intelligence in Construction
• Building Information Modeling (BIM)

Background:

• Prefabricated construction faces challenges in efficiency and economic performance due to limited resources.
• Optimizing multi-objective trade-offs in construction scheduling (duration, cost, resources) is complex.
• Existing scheduling models often lack adaptability and real-time feedback mechanisms.

Purpose of the Study:

• To develop an integrated intelligent optimization model for prefabricated building projects.
• To enhance construction efficiency and economic performance by optimizing assembly plans.
• To achieve a balanced compromise between construction period, budget cost, and resource conflicts.

Main Methods:

• Constructed an assembly semantic model using Building Information Modeling (BIM) and the BuildingNet dataset.
• Employed a multi-objective particle swarm optimization (MOPSO) algorithm with dynamic objective weighting.
• Integrated a Deep Q-Network (DQN)-based reinforcement learning strategy for real-time feedback and adaptive policy updates.

Main Results:

• The model generated feasible solution sets under varying objective weights for 100 assembly tasks.
• Achieved an average construction period of 85.2 days, a budget cost of USD 1.486 million, and minimal resource conflicts (<1.7 events).
• Outperformed rule-based, NSGA-II, and static MOPSO models in objective coverage, convergence speed, and solution diversity (HV=0.683, Spread=0.227, IGD=0.017).

Conclusions:

• The integrated model effectively supports dynamic multi-objective construction optimization in prefabricated projects.
• Combining semantic modeling, evolutionary optimization, and reinforcement learning enhances BIM practices for benefit-schedule-resource coordination.
• The approach offers a robust and adaptable solution for complex trade-off scenarios in construction management.