Graph-Based Generative Scheduling in Construction Project Management

Abstract: Factory-based modular construction resembles manufacturing systems, but production scheduling remains challenging due to design variability, product heterogeneity, and design–production coupling. Existing construction scheduling methods rely on predefined task structures and adapt poorly to design changes. This paper presents a graph-based generative scheduling framework that derives production plans directly from modular designs. A design graph is transformed through graph rewrite rules to generate tasks, resources, precedence constraints, and task durations. The resulting job shop scheduling problem is solved using constraint programming. Case studies demonstrate automated schedule generation, reveal resource bottlenecks, and support early-stage production capacity planning.