The placement of temporary emergency warehouses is critical for improving post-disaster relief responsiveness. However, earthquake induced road disruptions create uncertainty in travel times between warehouses and affected areas. This study investigates an uncapacitated stochastic facility location problem (SFLP) that minimizes the expected maximum travel time under stochastic road-delay conditions with a limited number of warehouses. A simheuristic approach combining a biased-randomized drop heuristic with a min-max penalty and Monte Carlo simulation is adapted to the 2022 Cianjur earthquake case in West Java, Indonesia. Unlike previous studies using synthetic disruption factors, this research calibrates lognormal travel-delay distributions through three-point PERT expert elicitation with the regional disaster management agency (BPBD). Results show that deterministic solutions fail to convert additional warehouses into stochastic performance, whereas the simheuristic configuration at the existing three-warehouse budget reduces the expected worst-case response time by 68.1%, underscoring the importance of warehouse planning for reliable humanitarian logistics.