Order Fulfillment and Replenishment Policies for Fully Substitutable Products
We consider a fully substitutable multi product inventory system in which customer demand may be satisfied by delivering any combination of products. A customer requesting a certain quantity, D may accept different combinations of products as long as its order of size D is fulfilled.We model the system as a Markov decision process and develop a approximate dynamic programming algorithm to determine order fulfillment and replenishment policies. In order to reduce the run time of the algorithm, we use two-layer neural network that iteratively fits a function to the state values and finds an approximately optimal order fulfillment and can order type replenishment policy.
KeywordsMarkov Decision Process Product Combination Replenishment Policy Markov Decision Process Model Inventory Replenishment
Unable to display preview. Download preview PDF.
- 2.Balintfy, J.L.:On a basic class of multi-item inventory problems. Management Science 10, 287–297 (1964)Google Scholar
- 5.Melchiors, P.:Calculating can-order policies for the joint replenishment problem by the compensation approach. European Journal of Operational Research 14, 587–595 (2002)Google Scholar
- 7.Thompstone,R.M. and Silver,E.A.:A coordinated inventory control system for compound poisson demand and zero lead time. Int.J.Prod.Res. 13, 581–602 (1975)Google Scholar