A Multi-Scenario Multi-Period Facility Location-Allocation Model and Algorithm for Pre-Disaster Planning

Abstract

Due to the diversity, periodicity, and complexity of disasters, government agencies and humanitarian organizations must engage in comprehensive pre-disaster planning. The facility location and emergency material allocation are particularly critical components with this planning process. Therefore, this paper designs a multi-scenario multi-period facility location-allocation model (MSMPFLA) for pre-disaster planning. The model also focuses on the diversity and periodicity characteristics of disasters by constructing various disaster scenarios to simulate emergency material allocation schemes in different rescue periods. To address this model, we propose a hybrid discrete crow search algorithm and material allocation algorithm (DCSA-MA). Experimental results indicate that the DCSA-MA significantly outperforms other algorithms in terms of solution quality, convergence rate, computation time, and stability. Through the Wilcoxon rank sum test, the DCSA-MA demonstrates superior optimization performance in majority of disaster scenarios. Consequently, DCSA-MA is an effective and stable method for solving the MSMPFLA problem. In comparison to traditional model, the MSMPFLA model offers more stable and feasible solutions, thereby validating its practicality and applicability in real-world scenarios. In addition, the impacts of penalty cost and the number of open facility on the MSMPFLA model are assessed in a series of sensitivity analyses, so as to inform decision-makers to a develop reliable emergency material allocation scheme.