A Hybrid Decision-Support Framework for Crisis-Induced Rescheduling in Infrastructure Projects: Integrating a Graph- Based Model with a Generative AI Interface

Abstract

This study presents a hybrid decision-support framework for crisis-induced rescheduling in a large-scale infrastructure project. Using a real-world case involving the insolvency of a critical architectural subcontractor, the study examines how project-continuation alternatives can be compared based on net financial gain, resource demand, and operational constraints. The deterministic component represents project activities and dependencies in a directed graph and calculates financial and resource indicators for the evaluated continuation alternatives. A generative artificial intelligence interface is then applied as an interaction layer over these deterministic outputs, enabling decision-makers to query and compare calculated scenarios in natural language. The empirical application illustrates how the framework can support the identification of a financially preferable continuation alternative and facilitate scenario comparison in a time-sensitive crisis setting. Since the evidence is derived from a single project structure and a single focal disruption type, the proposed framework is presented as a case-based methodological proof of concept that requires further validation across additional project and crisis contexts.