OPTIMIZATION OF DIMETHYL ETER (DME) SUPPLY CHAIN DAILY OPERATIONS IN SUMATRA WITH RAIL TANK WAGONS AND VESSELS USING MIXED INTEGER LINEAR PROGRAMMING
Indonesia import of LPG is currently up to 80% of its demand. To reduce its import dependency, PT X as the existing LPG distributor has made an agreement between PT Y as the producer and PT Z as the investor to produce DME in Tanjung Enim, South Sumatra, which serves as the substitute of LPG. PT X i...
Saved in:
Main Author: | |
---|---|
Format: | Final Project |
Language: | Indonesia |
Online Access: | https://digilib.itb.ac.id/gdl/view/72674 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Institut Teknologi Bandung |
Language: | Indonesia |
Summary: | Indonesia import of LPG is currently up to 80% of its demand. To reduce its import dependency, PT X as the existing LPG distributor has made an agreement between PT Y as the producer and PT Z as the investor to produce DME in Tanjung Enim, South Sumatra, which serves as the substitute of LPG. PT X intend to utilize their existing 4 depots in the 6 converted provinces to handle the supply and distribution of DME. PT X examine the possibility of using rail tank wagon (RTW) as a transportation mode because the plant and 2 out of 4 depots mostly have been interconnected by railway network infrastructure. From these 2 depots, DME is delivered using vessels to the two other depots. This will be a new distribution network for PT X. Therefore, PT X require optimal operational-level decision to deliver DME at a lowest possible cost while meeting the minimum coverage days which is set to 3 days at the depot levels.
The model is developed by using MILP to determine the daily scheduling of both RTWs and vessels to ensure sufficient demand for the multi echelon depots at the lowest possible cost. To reflect the dynamical aspect of daily operational scheduling, time is considered in the MILP model, as index in inventory and allocation variables. In this research, the model is solved by using two algorithms. Branch and bound algorithm which leads to optimal solution but can be used only for small scale problems because it takes a long time to solve. Therefore, another algorithm is developed to solve in a long planning horizon although it leads to suboptimal solution which can be classified into neighborhood search algorithm. Solution derived from developed algorithm is 10% greater than optimal solution. .
The results show that during 365 days of the planning horizon, PT X need 136 RTWs to Panjang Depot, 243 RTWs to Pulau Layang Depot, 70 vessels to Teluk Kabung Depot, and 93 vessels to Dumai Depot. This solution costs 19,136,994.20 USD or equal to 15.31 USD/MT meanwhile LPG current transport fee is 22.00 USD/MT. Average utilization in every depot is around 46% to 72% and the average of coverage days in every depot is 4 until 5 days.
|
---|