Scenario Based Approach To Decentralized Stock Distribution

Scenario Based Approach To Decentralized Stock Distribution: A Royal Netherlands Army case study

Koffeman, Rutger (TU Delft Mechanical, Maritime and Materials Engineering)

Duinkerken, M.B. (mentor)
Negenborn, R.R. (graduation committee)

Delft University of Technology

Marine Technology | Transport Engineering and Logistics

2018-07-30

Background: Currently the Royal Netherlands Army (RNLA) has a traditional three echelon supply chain consisting of suppliers that deliver to a warehouse and the warehouse that delivers to customers. However, the RNLA is interested in having its stock geographically spread for safety reasons. A solution for this would be to re-design the supply chain into a decentralized stock distribution design. This means that instead of keeping the majority of the stock at a RNLA controlled storage facility, the stock will be stored at the suppliers location and delivered when required. Updating the current RNLA supply chain network towards a decentralized product storage network might lead to a better performing network, but the downsides in case of an emergency response sit- uation are still unclear. This lack of knowledge could severely jeopardize the safety of the Netherlands, it’s citizens and also the employees of the RNLA. This research aims to answer the following question:
What RNLA supply chain design, with a focus on decentralized stock distribution, is the most effective in case of an RNLA emergency response situation?
Methods: Based on a literature research, interviews and a flow chart analysis the RNLA specific situation has been studied and translated into an extended supply chain design. Using the method of mathematical programming a multi-period, multi-commodity, multi-objective model has been created that describes the flow between multiple sources and sinks. Penalties are used as cost function to differentiate between desirable and less desirable product flows. By turning on and off different new supply chain functional- ities such as cross-docking, direct supplier deliveries and pop-up cross-docks different supply chain designs can be experimented with. User service level and decentralized stock distribution are used as key performance indicators. Fuzzy numbers have been used to account for uncertainty of demand. By utilizing both single and multi-objective objectives a complete picture of a specific situation can be created.
Results: Experimenting with different scenarios, which are combinations of supply chain setups and parameter settings, led to the classification of different supply chain designs based on the resulting objective values. Insight is created in decentralized prod- uct storage, early and late product flows and the maximum number of late days, based on the resulting product flows. It was noticed that the weighted-sum multi-objective method took less time to find the optimal solution, compared to the other methods. Conclusions and Recommendations: Using the supply chain classification, the best supply chain design can be chosen. Using the weighted-sum method, this can be done in an efficient manner. It is recommended to study additional key performance indicators such as the effect of costs. The effect of more advanced objective methods could lead to better results, especially the effect of penalty factors should be studied. Also, internal logistics and vehicle routing should be studied to improve the solutions. On top of that it is very important for the RNLA to accumulate the right kind and amount of data in order to be able to optimize their supply chain now and in the future.

Scenario based
Decentralised stock distribution
Emergency response
Royal Netherlands Army

http://resolver.tudelft.nl/uuid:2bd39ae9-7532-4b52-9659-97c8304a44cc

2019-07-30

Student theses

master thesis

© 2018 Rutger Koffeman