Mixed Integer Linear Programming as a Method for Evaluation of Location and Number of Buffer Warehouse in PT Petrokimia Gresik Distribution System (Case Study: West Java and Central Java)

The new policy issued by PT Pupuk Indonesia regarding the change of the responsible party for the procurement and the distribution of subsidized NPK fertilizers affected the subsidy allocation for PT Petrokimia Gresik, as the only company to conduct the procurement and distribution of the subsidized NPK fertilizer. It is now partially charged to the other holding members such as PT Kujang Cikampek, PT Pupuk Sriwidjaja, and PT Pupuk Kaltim. Also, all members of PT Pupuk Indonesia should face the reduction of subsidy allocation from Indonesia Government. PT Petrokimia Gresik also have to deal with Phonska V project plan in order to increase capacity of NPK fertilizer. Those policies are most significantlty impactful to the decrease of subsidy allocation in West Java. Besides, Central Java, a branch area of PT Petrokimia Gresik which has the biggest number of total capacity and fixed cost, should reevaluate the number and location of the buffer warehouse. This paper is aimed to explain the methodology using Mixed integer Linear Programming (MILP) in determining the optimal number and location of buffer warehouse in West Java and Central Java in response to the policies and the project plan. Firstly, the distribution system of NPK subsidized fertilizer are described. Next, previous researches about MILP and forecasting method applied in distribution system problem is explained. After that, related data and several conditions that are applied based on actual situation are shown. It can be deducted that MILP is an optimal method for the evaluation as it fits to the characteristic and condition of distribution system in PT Petrokimia Gresik.