복수의 공급자가 존재하는 공급사슬을 위한 공급계약의 최적화 모형과 해법

Abstract

Recently, many companies make considerable efforts for the management innovation in the challenging business environments with intensifying competition, short product life cycle and product variety increases. In order to gain competitive advantages over rival companies in these business environments, it is important to have abilities to supply competitive products in a timely manner and to respond to the change of the customer demand quickly. To have these abilities, companies can make a variety of supply contracts and try to maintain long-term relationship with suppliers, which can be frequently found in high tech industry such as memory and semiconductor. In this paper, we investigate quantity flexibility(QF) contract and capacity reservation(CR) contract. These contracts offer several benefits to supply chain members by providing flexibility to handle uncertain demand and also by permitting better capacity planning. We propose an open loop feedback optimal control(OLFOC) approach to find optimal value of contractual terms. We develop a linear optimization model and propose an implementation strategy by taking advantage of the rolling horizon procedure in order to realize the OLFOC. Numerical tests were performed to measure the efficiency of the proposed algorithm. We compare our algorithm with global optimal solution generated by the stochastic dynamic programming. The next tests were conducted under various environments simulating real world situations. It was shown that the cost savings can be up to 17% by the proposed algorithm. Afterwards, we analyzed the problem of the CR contract. A linear optimization model featuring several key component of CR contract was developed for a system with multiple suppliers. An implementation strategy utilizing the rolling horizon method was developed. Numerical tests simulating real world problems were done to verify the efficiency of the proposed algorithm. The results of the tests showed that the cost reduction could reach up to 15% when the proposed algorithm was employed. Finally, it is necessary to mention that the algorithms adopted in this paper cannot ensure global optimality of generated solution. In this regard, future research could be necessary to compare the performance of our algorithms to the one of a stochastic programming approach.