Fleet Management in Humanitarian Sector
研究人道主义组织在数据匮乏条件下如何确定最优车队规模与采购策略，以及现场车队管理政策，发现混合采购策略最优、车辆无需按任务类型分配、现有更换政策需调整。
Fleet management is a major concern for international humanitarian organizations because of (1) the magnitude of transportation related costs in humanitarian operations, second only to personnel cost and, (2) the pivotal role that transportation plays in the order fulfillment process. Humanitarian organizations face unusual operating constraints, which include working in areas with poor infrastructure, extreme environmental conditions as well as budget limitations. Most of the existing models derived from commercial supply chains are inapplicable in such a context. Therefore, a new set of tools and theories is required. This dissertation contributes to the development of such a new set of tools. It is composed of two parts that address two related questions in humanitarian fleet management: (1) how to determine the optimal fleet size and the optimal procurement strategy at an aggregate level and, (2) how to optimally manage an existing fleet the field level. Lack of data is the main challenge that prevents humanitarian organizations from adopting data-intensive models developed for commercial supply chains. Accordingly, the first part of this thesis studies how to determine optimal fleet capacity over time and how to minimize procurement costs for different demand profiles in the absence of detailed data. Contrary to conventional wisdom in humanitarian organizations, its findings show that a mixed policy of level and chase procurement strategies minimizes procurement costs and that a level strategy is the optimal approach to procurement in most humanitarian missions. The second part of the dissertation concentrates on fleet management policies at the field level. To optimize fleet performance and maximize demand coverage, humanitarian organizations implement policies to enhance the utilization of vehicles and minimize their physical depreciation. Through the analysis of a large humanitarian organization’s fleet in four representative countries (Sudan, Ethiopia, Afghanistan and Georgia) the results of this dissertation suggest that: (1) it is not necessary to assign different vehicles to specific mission types (2) all vehicles should be used following the same usage policy regardless of their mission type and, (3) the vehicle replacement policy implemented by most humanitarian organizations is not effective and needs to be reconsidered. Results also demonstrate that, on average, a utilization-depreciation trade-off does not exist and that a wellconceived fleet management policy can allow for both higher vehicle utilization and lower depreciation.