Live reoptimization of operating room schedules is required to cope with disruptions such as
emergency arrivals and deviations in surgery durations under strict time limits. The resulting problem
can be formulated as a large-scale Resource Constrained Project Scheduling Problem (RCPSP).
While exact optimization methods are attractive in this context due to their ability to provide strong
bounds, their application in live settings remains challenging because of the computational complexity
of large-scale instances. We propose an exact branch-and-price framework for live operating
room reoptimization in which pricing, pruning, and branching decisions are tightly integrated. The
pricing problem is solved by a bucket graph-based labeling algorithm supporting exact forward and
bidirectional search as well as multiple cycle-elimination strategies. We introduce two complementary
exact pruning mechanisms – surgical case blocking and problem-specific completion bounds –
and prove that they preserve LP optimality independently of the search direction and the chosen
cycle-elimination scheme. Extensive computational experiments and a simulation study demonstrate
that the proposed approach significantly improves computational efficiency and is suitable for live
planning settings.