DEMAND PREDICTION AND POLICY MAKE-TO-ORDER/MAKE-TO-STOCK MODEL DEVELOPMENT CONSIDERING PRODUCT LIFE CYCLE IN FASHION INDUSTRY

• Main Author: Apriana, Karina
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/51555
• Institution: Institut Teknologi Bandung

Demand prediction is crucial activity for production planning. Research on fashion demand prediction often aims to predict aggregate demand of a product family. This research tries to predict more detail product level which considering color variance. Demand prediction model for fashion industry is developed based on artificial neural network (ANN). Temperature, special days, store sales, location, discount percentage, price, population, style, materials, average of Indonesian expenses for fashion product, income, and e-commerce user trend are among variables that is used in the model. The result showed a good accuracy and can be applied as a new product prediction. In this research, the result that is obtained from ANN is being a basic for MTO/MTS policy planning. Researches on hybrid MTO/MTS production planning often tries to decide order acceptance or rejection because they focus in MTS or MTO perspective. This research developed model to determine MTO or MTS production policy in a period without certain perspective. Policy MTO/MTS is applied in 4PL company who manage fashion industry. Markov decision process (MDP) model is developed to obtain optimal policy considering product life cycle. The optimal policy showed decision to make for an observed state. The decision varies for particular inventory level, MTO order, and life cycle stage by considering lead time, minimum order quantity, and maximum amount of MTO order allowed in the system