PyRebuild: A Python-Based Simulator for the Dynamic Post-Earthquake Housing Reconstruction Problem

Post-earthquake housing reconstruction faces a key challenge: damage assessments are reported gradually over time, while conventional scheduling methods assume that complete information is available at the start. This discrepancy can lead to inefficient resource allocation and delays in recovery. We introduce PyRebuild, a Python-based simulation framework that models the temporal dynamics of damage report arrivals and dynamically allocates reconstruction resources. The framework tests three scheduling algorithms (longest job first, shortest job first, and random assignment) under static and dynamic conditions. Using 2018 Lombok earthquake data, we find dynamic scheduling yields lower root mean square error between predicted and observed recovery timelines in five of seven regions, while static methods perform marginally better in two regions. Results indicate optimal scheduling depends on local damage patterns, highlighting the need for adaptive post-disaster reconstruction strategies.