Multi-Echelon Inventory: Impact of Varied Ordering Policies on Realized Service Levels
Open Access
- Author:
- Brady, Stephan Paul
- Graduate Program:
- Business Administration
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 01, 1999
- Committee Members:
- Joseph L Cavinato, Committee Member
Lisa Williams, Committee Chair/Co-Chair
Alan J Stenger, Committee Member
Jose Antonio Ventura, Committee Member - Keywords:
- Supply Chain
Simulation
Ordering Policies
Inventory Control - Abstract:
- Wholesalers face an environment where different retail establishments may follow different ordering policies. Typically, wholesalers do not consider the individual retail establishment’s ordering policies when setting and computing their service levels. This research looked at whether there was a negative impact on retailers due to interactions between ordering policies. Current research has focused on serial relationships in multi-echelon inventory problems. This research extends existing research by assessing the impacts of differing ordering policies upon service level performance for each retail level establishment in a parallel multi-echelon inventory environment, for slow-demand items. Conducted through the use of simulation, one model was developed representing three ordering policies: Periodic Review, Continuous Review, and an Instantaneous Reorder Policy. Each policy was isolated, and the models run to get a baseline of performance. Finally, the complete model was run, and comparisons were made between the performance of the policies alone, and interacting. Runs were conducted over a range of customer demand rates and target service levels. Lead times and customer demand rates were exponentially distributed. Of interest is whether the achieved service level at least meets the target service level for the individual run. This research focused on four research questions. This research concluded that the three reordering policies generally performed as well as, or better, than the theoretical target service levels, both for when they were run independently, and for when the policies were allowed to interact. There was generally no statistical difference between the performance of identical policies when isolated and when run concurrently. Additionally, it was found that there exists a statistical difference between the policies when interacting. Typically, the periodic review policy’s achieved service levels was lower in comparison to the other two policies under conditions of slow demand, and was higher than the other policies as demand approached one per day. While some policies outperform others, each policy meets or exceeds the target service level, suggesting that other considerations may be more appropriate to decision makers. This research concludes that inventory reordering decisions should continue to be made using economic and other factors. Further research should look at other ranges and distributions of demand, as well as lead-time. Additionally, other influences on the supply chain may be integrated into the model. Finally, analysis of the cycle stock for each ordering policy may be of greater value in assessing costs reductions available through alternative ordering policies.