A stochastic dynamic programming model for multi-site capacity planning in TFT-LCD industry

Abstract

This research focuses on the multi-site capacity planning problem in TFT-LCD industry under stochastic demand. TFT-LCD industry becomes a critical industry in Taiwan and has the following characteristics: (1) complex product hierarchy and product types; (2) coexistence of multi-generation manufacturing technologies in the multi-site production system; (3) rapid changing and stochastic markovian demand. In this complicated environment, the capacity planning which balances demand and supply is encountered by TFT-LCD industry. It is to simultaneously make a robust capacity allocation plan and capacity expansion policy across multiple sites against stochastic demand. In addition, for the TFT-LCD manufacturing, the demand situation follows Markov properties in which the correlations of the demand variations in the consecutive periods are high and the demand status in next period is stochastically determined by the present one. In this regard, two-stage stochastic programming traditionally neglects dynamic capacity adjustments in response to new available demand information. Therefore, this study constructs a stochastic dynamic programming (SDP) model which considers many TFT-LCD practical characteristics and constraints to generate a dynamic capacity planning policy as the demand in each period is revealed and updated. In the end, a numerical example is illustrated to prove the feasibility of the proposed model.