The COVID-19 pandemic prompted the implementation of “bubble groups” in schools to minimize transmission by limiting interactions among students. However, this approach overlooked latent sibling interactions across classrooms, reducing its effectiveness. This study introduces the sibling rewiring problem, a network-based framework to optimize student assignments and minimize cross-group connections facilitated by siblings. Using synthetic datasets, we evaluated three strategies—random assignment, bubble groups, and a heuristic optimization algorithm. The heuristic achieved superior network fragmentation, significantly increasing the number of connected components compared to the other strategies. Despite its effectiveness, challenges remain, including balancing group sizes for larger family distributions and addressing logistical constraints. Future work should explore advanced optimization techniques and multicriteria objectives to enhance applicability in diverse contexts.