How to provide a humanitarian warehouse location model with data - An assessment of scenario creation and data availability at the United Nations High-Commissioner for Refugees (UNHCR)

Sammanfattning: At the end of 2014, 38 million people around the world had fled their homes due to man-made disasters such as war, armed conflict and violence. To save the lives and alleviate the suffering of these refugees and displaced is the purpose of Humanitarian Logistics. Preparedness is of key importance in humanitarian organisations as it is a tool to reduce the impact of coming events. One common preparedness measure is prepositioning, i.e. to keep stock ready in anticipation of a disaster to ensure quick dispatch after its onset. Prepositioning is one of the measures UNHCR, the United Nations Refugee Agency, is engaged in to increase its general preparedness. The volumes stored in global stockpiles allows the organisation to assist up to 600 000 refugees with emergency relief within 72 hours. To deliver to this high standard, the geographical placement of the global warehouses of UNHCR is paramount. Currently the organization has seven global stockpiles located in Cameroon, Denmark, Dubai, Ghana, Jordan, Kenya, and Tanzania. To ensure the optimality of the warehouse distribution the UNHCR has developed a humanitarian facility location model in collaboration with Lund University, BI School in Oslo and North Eastern University in Boston. This model allows the organisation to computationally determine where each stockpile should be placed and the aim is to implement the model by Q3 2017. Before that can happen existing barriers and challenges need to be identified and overcome, which is the purpose of this thesis. This was carried out by performing an in-depth case study at the regional headquarters in Budapest, Hungary, including a survey at the bi-annual Logistics Cluster meeting and a field visit to Amman, Jordan. The thesis found eight challenges associated with the implementation of the model regarding scenario creation and data availability that can be classified into four areas of improvement that are key to overcome them. The current upgrade of the ERP system provides a golden window of opportunity to mitigate these challenges. The four areas that need to be improved are increased internal communication, enhanced preparedness, staff training and increased functionality of the ERP system. Increased communication and enhanced preparedness has to do with that the contingency plans for preparedness are not shared or aggregated within the organisation and that much information is stored outside the ERP system in offline spread sheets and communicated via email. This is due to suboptimal staff behaviour in terms of data input in the systems and choice of communication channels. In addition the ERP system of UNHCR does not contain sufficient functionality to feed the model or share information throughout the organisation. Only 35% of the functional requirements of the model can currently be satisfied. Another issue of the ERP is that the information that is stored is not sufficiently detailed which greatly hampers its usability. Mode of transport is a prime example where 67% of all logs are generically classified instead of stating the used one. Further standardising the core relief items shipped could also enable automatization of delivery information such as weight and volume of shipments, something which is currently done manually. To feed the model with future scenarios the research team is recommended to create a baseline scenario by using the UNHCR Population statistics and the dollar worth of the standard aid provided. Sensitivity analysis can then be used to find the best and worst case scenario for the supply chain. When the new functionality of the ERP system and the demand application Demantra is implemented it will become possible to extract future demand scenarios directly from the systems of UNHCR. Until then this method will provide future scenarios necessary to run the model. This thesis, being one of the first to evaluate scenario creation and data availability for humanitarian facility location models, has also found a need for future research when it comes to conflict disasters and a greater need for transparency in the field of Humanitarian logistics. Most literature deals with facility location in the context of natural disasters. Little research exists to know if this is transferable to man-made disasters in a good way. Also, most articles are not sufficiently transparent in how the parameterisation of the supply chains was carried out or how the scenarios were created. However, the need for this kind of decision support tools will only become more important as more disasters occur and the current number of displaced people in the world is the highest in the history of mankind.