A study on the assemble-to-order system with common component

Component commonality has been widely recognized as an important way to achieve product variety at low cost. This topic has drawn lots of attention from researchers and most results are built primarily on static models (single-period models or multi-period, periodic-review models with zero lead times) and their results typically show that using common component can be beneficial and significant when the number of products increases, demand correlation decreases and the degree of commonality increases[1][2]. Unlike most of the studies on the component commonality problem before, we study a multi-product ATO system which has stochastic component lead times is positive and demand for the product takes place continuously over time according to an independent Poisson process. We formulate the problem as a Markov decision process and by conducting a series of numerical studies we find that some properties exist in the static models do not hold in dynamic circumstance anymore.