TY  - CONF
TI  - A multi-agent data-to-model workflow for make-ready assessment
C1  - Singapore, Singapore
C3  - Proceedings of the 34th Annual Conference of the International Group for Lean Construction (IGLC 34)
SP  - 238
EP  - 249
PY  - 2026
DO  - 10.24928/2026/0266
AU  - Qi, Ruixuan
AU  - Xu, Jinying
AU  - Zheng, Ruiyan
AD  - Graduate student, National University of Singapore, Singapore, e1538237@u.nus.edu
AD  - Assistant Professor, Department of the Built Environment, National University of Singapore, Singapore; jinying.xu@nus.edu.sg, orcid.org/0000-0001-9589-6396
AD  - PhD student, Department of the Built Environment, National University of Singapore, Singapore; zhengruiyan@u.nus.edu, orcid.org/0009-0009-8421-0614
ED  - Hamzeh, Farook
ED  - Poshdar, Mani
ED  - Garcia-Lopez,, Nelly P.
AB  - The Last Planner System (LPS) is widely recognized as an effective lean approach for improving workflow reliability. Although the theoretical framework of LPS has been well established, its practical implementation remains challenging in real projects. A major barrier lies in initiating and maintaining Lookahead planning, which is intended to connect phase scheduling with commitment planning. Current practice often shows gaps, including lack of detailed data required for Make-Ready assessment in Lookahead planning.
This paper proposes a multi-agent, data-to-model workflow that lowers the entry barrier to Lookahead planning by extracting project-specific Lookahead rules and readiness assessments directly from commonly available progress schedule data. Instead of training a generalized predictive model, the workflow coordinates specialized agents to sequentially normalize tasks, infer dependencies, reconstruct production states, and support readiness scoring with an interpretable learning component.
The workflow is demonstrated using real data from a high-rise building project. The case suggests that meaningful Make-Ready assessments can be derived from standard project schedule data. It may help practitioners reduce reliance on subjective judgment and fragmented information in Lookahead planning, while offering a practical method to push near-term work commitments further in LPS adoption.
KW  - Last Planner System
KW  - lookahead planning
KW  - make-ready assessment
KW  - multi-agent systems
KW  - data-to-model workflow.
PB  - 
T2  - Proceedings of the 34th Annual Conference of the International Group for Lean Construction (IGLC 34)
DA  - 2026/06/22
CY  - Singapore, Singapore
L1  - http://iglc.net/Papers/Details/2555/pdf
L2  - http://iglc.net/Papers/Details/2555
N1  - Export Date: 19 June 2026
DB  - IGLC.net
DP  - IGLC
LA  - English
ER  -