الگوریتمی برای طراحی سیستم تک مرحله ای کانبان تطبیقی

The traditional kanban system with fixed number of cards does not work satisfactorily in unstable environment. In the adaptive kanban-type pull control mechanism the number of kanban is allowed to change with respect to the inventory and backorder level. It is required to set the threshold values at which cards are added or deleted which is a part of the design. Previous studies used the local search method to design the adaptive kanban system. In this paper Genetic Algorithm- and Simulated annealing-based heuristics are developed and used to set the design parameters of adaptive kanban system. The numerical results indicate that simulated annealing based heuristics produces better solution with improved computational efficiency.

In pull-production system kanbans are used as production orders. In the Traditional Kanban System (TKS) the number of cards used in a manufacturing process (MP) is kept constant. Hall (1983) proved that TKS is successful in production environment with stable demand and lead time. Di Mascolo, Frein, and Dallery (1996) developed an analytical method for performance evaluation of TKS. Wijngaard, 2004 and Karaesmen et al., 2004 used inventory control policies which resulted in, significant cost savings in the TKS through inventory reductions and improvement in customer service. Liberopoulos and Koukoumialos (2005) have presented a simulation model of TKS in which the effect of advanced demand information is analyzed. However, in unstable environment the TKS does not work satisfactorily. Philipoom, Rees, Taylor, and Huang (1987) and Rees, Philipoom, Taylor, and Hwang (1987) investigated the key factors that affect the number of kanbans in the system. Few authors have discussed systems in which the number of kanban cards in use is adjusted according to the status of the MP. Such systems are called flexible or adaptive kanban system (AKS) (Valerie Tardif & Lars Maaseidvaag, 2001). In this paper, based on a Markovian model of kanban system use of meta-heuristic are explored to estimate the design parameters of the AKS. The performance of such system is compared with TKS and AKS reported in literature.

Traditional kanban system with fixed number of kanban performs unsatisfactorily under unstable demand. Adaptive kanban is proposed to over come this problem. AKS allows flexibility of varying number of cards used based on certain threshold values. Previous studies used the local search method to design the adaptive kanban system. In this work application of meta heuristics for the design of single-stage adaptive kanban system is explored. The objective is to minimize the cost associated with the inventory and backorder demand. GA- and SA-based search models are developed and analyzed. It is found that the SA-based search gives better value for the objective function coupled with considerable improvement in computational effort. The objective function value is improved up to 13.12% while CPU time is reduced up to 92.04%. Also it is found that in some cases the upper limit of cards used it self is less than that of TKS. Hence it is concluded that in the design of AKS, use of SA gives better results with less computational effort. In this work AKS is designed to handle single part only. However this can be used as a building block to design AKS to handle multi product in future. In addition, extending this work for multi-stage AKS can also be considered in future.