University of Twente Student Theses

Login

How VMI can be successful in gas distribution: a solution methodology for the inventory routing problem in gas distribution

Hulshof, P.J.H. (2008) How VMI can be successful in gas distribution: a solution methodology for the inventory routing problem in gas distribution.

[img]
Preview
PDF
25MB
Abstract:Introduction: This research is done at the department Oil, Gas, and Chemicals at ORTEC, a planning software company. We study the distribution of gas to commercial and residential customers that are not connected to a network of gas pipelines. These customers receive gas deliveries under a Vendor Managed Inventory (VMI) contract, which gives gas companies the flexibility to determine when and what volume to deliver, and what routes to choose. The decision problem that is associated with VMI for a large set of customers is the Inventory Routing Problem (IRP). Additionally, gas companies want to control the effects of the large seasonal peak in gas demand, to use the available resources efficiently. This research assumes customer usage to be deterministic, and we develop a solution for a region with multiple depots and vehicles with varying capacity (heterogeneous fleet). - Objective: To design a solution methodology to minimize distribution costs in the IRP for gas distribution, and mitigate the seasonal peak in customer deliveries. We propose a solution methodology that increases the volume per kilometre, since it is an important indicator of distribution costs. Additionally, we balance the delivery volume in the planning period, to use the available resources efficiently. To mitigate the seasonal peak, we balance the delivery volume over a relatively long period, so that the workload is more equally divided over the year. - Solution: An algorithm is developed to select a certain delivery day in the planning period for every customer. The algorithm focuses on finding delivery days for customers that can receive a relatively large delivery compared to the customer’s capacity, while minimizing total travel distance. Customers that require a delivery in the planning period must be planned, and the customers that do not require a delivery in the planning period are planned according to the impact of the delivery on total travel distance and on future planning periods. The delivery volume is balanced according to the available vehicle capacity to smoothen the delivery volume over the course of a year and to efficiently use the available workforce. - Results: Actual delivery data froma large gas company are used to test the algorithmin a planning period of seven days. The computational experiments show that the solution increases the delivered volume per kilometre by more than 21%. The delivery volume is balanced on the short-term and long-term, and is responsive to changes in the vehicle capacity in the planning period. Conclusions The solution decreases the costs for gas distribution, and requires an acceptable computation time. The long-term balance in delivery volume flattens the customer delivery curve, and thus helps in mitigating the seasonal peak.
Item Type:Essay (Master)
Clients:
ORTEC
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:http://purl.utwente.nl/essays/59271
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page