İnsani operasyonlarda çok sınıflı talep ve finansman belirsizliği altında dinamik envanter tahsisi

This study develops a dynamic inventory control framework for humanitarian aid allocation under multi-class demand and uncertain funding. We consider a central planner responsible for distributing a critical aid item to camp based and surrounding urban beneficiaries. Demand arises stochastically from two groups: internal (camp) demand, which incurs time sensitive deprivation costs when unmet, and external (urban) demand, which incurs fixed referral penalties when unsatisfied. Inventory replenishment is constrained by a known total funding amount that arrives in uncertain installments over a finite horizon. To approximate real time fulfillment, we employ fine grained intra period discretization, enabling dynamic rationing decisions at the micro period level. The problem is formulated as a finite horizon stochastic dynamic programming that integrates stochastic funding arrivals, continuous time demand realization, and threshold based rationing policies. We examine both linear and exponential time sensitive deprivation cost structures. Numerical experiments and structural analysis demonstrate the cost effectiveness and adaptability of dynamic policies under varying funding regimes. The proposed framework provides rigorous analytical foundations and actionable insights for resource constrained humanitarian supply chains.