Synchronized out-of-order execution for smart prefabricated construction

Abstract

Modular Prefabricated Construction (MPC) is increasingly seen as a promising solution to the problems of inefficiency, labor constraints and inconsistent quality in traditional construction. By shifting a significant amount of construction activity from on-site to off-site, MPC achieves higher efficiency and greater sustainability. However, MPC is still facing challenges such as fragmented workflows, lack of synchronization in the production and logistics phases, and uncertainties in the supply of materials, labor allocation and module transportation. These problems significantly hamper operational efficiency.

To address these challenges, a Graduation Intelligent Manufacturing System (GiMS) was developed to enable visibility, traceability and adaptability of prefabricated building module production. Inspired by the ticket-based queuing approach in graduation ceremonies, GiMS structures production and logistics activities through job tickets, setup tickets, and operation tickets to facilitate multi-stage synchronization in the planning, scheduling, and execution levels. A multi-stage adaptive decision-making mechanism was proposed, including Internet-of-Things (IoT)-enabled real-time data collection, spectral clustering for planning, and dynamic sequencing for real-time execution. Numerical studies demonstrate that GiMS can significantly reduce idle time at the mold table under various uncertainties.

In addition, the thesis investigates Prefabricated Off-site Fit-out (POF), an important but often under-explored phase of MPC. A GiMS-based synchronization mechanism was proposed to address the coordination of fixed-position operations and intralogistics in a dynamic fit-out environment. With real-time data acquisition and flexible resource allocation strategies, the system effectively manages uncertainty and improves production-intralogistics synchronization performance.

Further, to manage more complex and interdependent operations in the prefabrication yard, an Out-of-Order (OoO) synchronization mechanism was introduced. The mechanism draws inspiration from the Central Processing Unit (CPU) architecture to dynamically prioritize jobs and operations based on real-time situations. A spatial-temporal analytics model is developed to locate uncertainties and disturbances in off-site fit-out operations. The evaluation in the numerical study validates that the OoO synchronization mechanism improves resource utilization, minimizes job latency, and ensures a smoother workflow under varying operating conditions.

Finally, a prototype system enabled by Cyber-Physical System (CPS) was developed to integrate several subsystems throughout the MPC value chain, including off-site factories, prefabrication yards, on-site assembly, and prefabrication logistics. A physical model based on a real hospital construction case in Hong Kong demonstrated the feasibility of the proposed solution. The prototype utilizes real-time monitoring, data-driven decision-making, and intelligent coordination to improve resilience and scalability.

This research makes several innovative contributions. Firstly, a ticket-based smart manufacturing architecture is tailored for MPC. The OoO synchronization mechanism was proposed for construction production environments. Secondly, a cyber-physical prototype is developed that connects a theoretical model with a practical implementation. By integrating real-time sensing and adaptive scheduling, the thesis lays a solid foundation for the development of a prefabrication yard. This important node in the MPC life cycle can achieve resilient, responsive, and embody the principles of Industry 4.0. The proposed framework provides theoretical insights and practical ways to transform MPC into a smart and scalable ecosystem.