Bi-objective location allocation model for flood relief centres: a case study in Kuala Kuantan / Noridayu Mah Hashim

The frequency of natural disaster has been increasing over the years, which has resulted in the loss of life, damage to properties, and destruction of the environment. Compared to other natural disaster, floods are the most significant natural hazard in Malaysia that has affected thousands of people. Thus, a proper planning in locating relief centres especially at flood-prone areas is crucial in order to have good management in evacuating victims quickly and efficiently to avoid further worsening incidents. However, during these recent floods, several relief centres were experiencing problems when most of the flood victims were required to relocate to other relief centres after the existing shelters were affected by flood water. Furthermore, the relief centres became congested due to the receive of a large number of flood evacuees over the capacity provided. This study attempts to develop Bi-Objective location-allocation models for the allocation of flood victims during a flood event using the 'split is allowed' and 'split is not allowed' procedures. The assessment of the impact rainfall amount on the flood occurrence has been conducted. Two basic location-allocation models with two different objectives were studied: the P-median and Location Set Covering Problem (LSCP) models. Several solution procedures have been applied by considering the un capacitated and capacitated constraints in the Bi-Objective models. The allocation of flood victims in Kuala Kuantan was studied using the model, in addition to analyses on the past and current location decisions. These models were solved using Excel solver, sensitivity analysis, Pareto analysis, and Visual Studio 2019. The finding has been compared with the actual allocation process conducted by the Department of Social Welfare (DSW). A few recommendations have been highlighted to disaster management, including the changes in flood victims' allocation pattern and enhancement of the existing allocation process to be more systematic. On the other hand, the results of this study will also contribute to avoid congestions at relief centres during an allocation process since the flood victims are alert to which relief centres they need to move to during a disaster. The assessment of rainfall amount has indicated that the actual flood event occurred in prediction period. Capacity analysis has also been conducted to solve the congestion issues at relief centres during a disaster event. Moreover, the classification of prone areas (river, inland, and coastal) was found capable to enhance the existing allocation procedure.