Dynamic resource allocation for humanitarian relief chain coordination integrated with volunteers

Abstract

Each disaster situation has its own specific resource requirements and structures. In the event of a disaster, the number of injuries and lives lost, financial needs, and aid requirements vary based on the type of the disaster. Furthermore, the scale of the disaster, its location, and the socio-economic level of the affected region affect required resources and the actors to participate in relief operations and coordination structure. Volunteers are usually the early responders to the disaster, and need to be employed, as mass casualty events cause a sudden demand surge, and the governmental resources are usually not enough to satisfy all human resource requirements. In this study, a multi-objective dynamic resource allocation model is developed for search-and-rescue and first aid activities under uncertainty with a focus on saving human lives by prioritizing urgent tasks, where the volunteers are employed. A detailed framework covering these tasks and required resources is presented. We considered search-and-rescue units composed of team members from certain fields of expertise and volunteers as human resources. Moreover, the use of different types of renewable and nonrenewable resources required during these tasks and pre-positioning and distribution of these according to needs within the model were planned. In the model, it is aimed to minimize the total unmet human resource demand, the number of resources transferred between regions, and the total unmet renewable and non-renewable resource demand. Hence, a three- objective stochastic programming model is formulated. A real case study was conducted to address resource allocation after the expected major Marmara Earthquake for a district of Istanbul. Demand assessment was made considering the number of buildings expected to be damaged or destroyed, the classification of victims according to the NATO triage system and the number of victims expected to suffer with minor or severe injuries. To meet this demand, officially appointed rescue units and trained volunteers expected to receive training at predetermined levels of classified volunteerism in different professions and spontaneous volunteers are planned to work together in coordination. As the volunteers are usually the first responders, their consideration of resource allocation is invaluable. This is the first study that simultaneously places and dynamically allocates renewable and non-renewable resources and human resources, including the official rescue units and volunteers, for the search-and-rescue and first-aid tasks. Incorporating volunteers' assignments to the search-and-rescue and first-aid teams, considering their expertise levels and prior training, and quitting behavior are unique contributions to our study. The Augmented Epsilon-Constraint-2 (AUGMECON2) Algorithm was employed where necessary iterations were performed in the IBM ILOG CPLEX software environment to obtain the Pareto optimal solutions. Results were analyzed with sensitivity analyses. Besides, the model was solved based on the equity policy to minimize the maximum unmet demand among different regions. The outcomes of this research offer some administrative insights, in which ideas for resource allocation planning were suggested, considering both human and material resources. Most of the unmet human resources demand occurred in the first period since additional resource deployment and interregional transfer of human resources could not be achieved. The population density and the high number of casualties in some regions cause the unmet demand in these regions to be higher than in others. Although targeting the equity policy in total unmet demand among regions in the model is expected to provide equitable results, the total unmet demand, in this case, was much higher than without aiming for equity. As the model's results show, the lack of even one resource will prevent the completion of the task. In cases where some of the resources are used together, a communication system can be used to synchronize them. In the same way, a system that dynamically transmits the number of available resources in the regions to the center will be extremely beneficial in terms of ensuring the management of the process with optimum resources by preventing extra resources from coming to the regions. This requires pre-planning for healthy data flow from disaster areas to transmit the current situation in the region to the system. Integrated decisions for pre- and post-disaster stages are crucial for potential disaster scenarios, as they help determine the expected resource requirements. Improper pre-disaster decision-making will affect post-earthquake decisions, leading to a shortfall in meeting demand.