Dynamic allocation of kanbans in a manufacturing system using perturbation analysis

In this paper, an application of perturbation analysis techniques to the problem of dynamic allocation of buffering capacity in a multipart-type manufacturing system is studied. The problem is of extreme interest, among others, in kanban systems. The first contribution of the paper consists of a formulation of the generation function for the case of buffer size perturbation. The second contribution consists of an online control strategy, aimed at dynamically varying the buffering capacity of the working centres, in order to keep the performance of the system as high as possible, despite unknown disturbances entering the system. The proposed control strategy has been tested by means of simulation on a multipart-type kanban system. The control action consists of dynamically varying the number of kanbans associated to each part-type, in order to counteract the effect of unknown changes in the product-mix on the system throughput