DETERMINATION OF RAW MATERIAL PURCHASING FREQUENCY AND SCHEDULE CONSIDERING FLUCTUATING FOREIGN EXCHANGE RATE, CONTRACTED PRICE, AND DELIVERY LEAD TIME

• Main Author: Britania, Rizka
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/72573
• Institution: Institut Teknologi Bandung

This research presents problem on determination of optimal raw material purchasing frequency and schedule considering fluctuating foreign exchange rate, contracted price, and delivery lead time. This problem happens in companies in Indonesia which make transactions with different currency, and have quantity contract with supplier, such as PT XYZ, because every transactions in Indonesia has to use Indonesian Rupiah (IDR) as rule issued by Bank Indonesia No. 17/3/PBI/2015. The aim of this research is to give optimal purchasing frequency and schedule minimizing total cost of inventory. Purchasing quantity is based on total demand between consecutive schedules. Contributions of this research are on improvement of inventory model in terms of application of fluctuating foreign exchange rate, new contracted price, lead time, and algorithm to obtain discrete optimal purchasing schedule. There are three steps needed to solve the problem; forecasting foreign exchange rate (Rp/$), forecasting new contracted price ($/kg) from supplier, and solving the inventory model. Forecasting foreign exchange rate (Rp/$) is solved using Artificial Neural Network (ANN) method with Design of Experiment on number of input nodes and hidden nodes. Forecasting new contracted price ($/kg) is solved by time series forecasting method, ARIMA as it gives better performance than Geometric Brownian Motion in this case. ARIMA model that fits the data is ARIMA (1,1,0) with coefficient 0,8026. These forecasts are used as input in inventory model to obtain optimal raw material purchasing decisions. The model is solved with analytic and numeric method. Performance of model is tested with actual purchasing from PT XYZ on March 2017 case. The result shows the differences of frequency and schedule between the model and the actual condition with 0,41% saving. Sensitivity analysis shows that the developed model and algorithm is not sensitive to the changing of holding cost fraction and ordering cost but sensitive to demand and exchange rate (Rp/$) mean value.