A COOPERATION MODEL BASED ON VENDOR MANAGED INVENTORY (VMI) FOR A SINGLE NGO AND A SINGLE VENDOR IN FULFILLING PERISHABLE DISASTER RELIEF PRODUCTS DURING VOLCANIC EMERGENCY RESPONSE

• Main Author: Prasetyo, Ikhsan
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/83706
• Institution: Institut Teknologi Bandung

Indonesia is one of the countries with geographical characteristics significantly more susceptible to natural disasters. One of the most significant threats is the potential for volcanic eruptions, due to the country’s location within the Ring of Fire. As a nation with over 120 volcanoes, approximately 40% of these volcanoes remain insufficiently monitored. Nonetheless, the Indonesian government has recognized the critical importance of volcanic risk management, incorporating it as a key priority within the framework of its National Development planning strategy. A major challenge in disaster management is the effective handling of humanitarian logistics, as 65% of the allocated budget is typically devoted by NGO in procurement activities. This issue becomes even more complex when the required aid comprises perishable products with limited shelf life. Furthermore, the unpredictable nature of disaster occurrences creates significant volatility in both the supply and demand for humanitarian relief. This volatility is exacerbated when aid provision is predominantly reliant on post-disaster donations, which frequently arrive after the initial impact. As a consequence, the duration of the emergency response phase is often extended beyond optimal timeframes, potentially compromising the efficacy of relief efforts. This research develops a model for managing the procurement of perishable products with fixed shelf-live during the critical first 72 hours of volcanic disaster emergency response, based on the Vendor Managed Inventory (VMI) concept. The proposed model focused on a two-echelon Vendor Managed Inventory (VMI) system between an NGO and a vendor. The proposed model was conceptualized and developed with the primary objective of ensuring expeditious and efficient distribution of perishable commodities, particularly foodstuffs, in the immediate aftermath of volcanic events. This strategic approach is designed to facilitate a rapid and effective emergency response, thereby potentially reducing the duration and impact of the crisis period. In this study, the vendor is assumed to face two types of demand: direct customer demand and disaster-related demand, each with distinct distribution characteristics. The demand fulfillment process involves inter-channel transfer, meaning these two demands are interrelated. Specifically, a product can potentially fulfill either direct customer demand or disaster demand, depending on whether a disaster occurs in each period. If a disaster occurs, all products are first allocated to meet disaster demand, with any remaining inventory then used to fulfill direct customer demand. iv Two models are developed: a mathematical VMI model and a simulation VMI model. The mathematical VMI model is designed to determine the optimal values for Q (quantity ordered) and F (order frequency), while the simulation VMI model, based on the vendor's event diagram, evaluates the performance of these optimal solutions against changes in random parameters throughout the planning period. The two models were formulated with the primary objective of mitigating the multifaceted risks encountered by vendors and non-governmental organizations. These risks encompass product obsolescence, aggregate inventory holding costs, and shortage-related expenses stemming from unfulfilled demand, whether originating from direct consumers or disaster-induced requirements. The models' design aims to optimize resource allocation and enhance operational efficiency in both routine and crisis scenarios. To address the research problems, a multi- method stochastic simulation (MMSS) approach is employed, combining several techniques including Markov Decision Process (MDP), Bernoulli trials, random sampling, and Monte Carlo simulation, to provide a more comprehensive framework for decision-making. Decision-making processes are predicated on the optimal scenario that minimizes the aggregate risks faced by non-governmental organizations and vendors. Two distinct scenarios are considered in this analytical framework. The first scenario posits the application of optimal conditions for quantity (Q) and frequency (F) under the assumption of no disaster occurrence during the planning horizon (R = 0). Conversely, the second scenario contemplates the implementation of optimal conditions for Q and F under the presumption of a disaster event within the planning period (R = 1). Subsequently, both scenarios undergo rigorous evaluation to assess the potential risks associated with their respective selection as the preferred decision pathway. The parametric criteria for scenario selection are fundamentally based on the minimization of a tripartite cost structure: product obsolescence expenses, inventory holding costs, and shortage- related expenditures. The latter encompasses both unfulfilled direct consumer demand and unmet disaster-induced requirements. This comprehensive approach ensures a holistic consideration of potential financial and operational implications in the decision-making process.