Optimization of Backorders at The Warehouse Node of Single-Node Single Product (SNSP) Supply Chain, Running Under Model Predictive Controller Framework, In A Constrained Environment
Abstract
In this article, a single-node and single-product (SNSP) supply chain model, is investigated to reduce backorders at the warehouse node of the network. First, the stochastic demands for two months are created using random number generators. After that, a mathematical model of SNSP is constructed using mass balance equations at the warehouse node. During modeling, care has been taken to handle five types of supply chain disruptions like; the delay in shipping from the manufacturer to the warehouse, damage of inventory during transportation, damage of inventory in the warehouse itself, random demand pattern, and cancellation of backorders. As supply networks are dynamic networks; Inventory, shipping, and delivery of the product at different sampling time instances must be in tune with each other to fulfill the stochastic demand at the dealer's node. Today’s supply networks need dynamic optimization for their smooth running to nullify the effect of aforesaid disruptions. The model predictive controller existing in MATLAB is found suitable to handle these disruptions as compared to other techniques like a neural network, fuzzy logic, etc. because it can incorporate network delays, uncertainties, and constraints, in an efficient manner. A MATLAB code (program) is prepared to capture the mathematical model in MPC (Model Predictive Controller) framework. The MPC framework provides the dataset of running the SNSP network for sixty days. This dataset and MATLAB code may be helpful for supply chain personnel to minimize the backorders at the warehouse node and to maximize customer satisfaction under constraint and competitive environment in the present scenario. The supply chain head can look into the future about the running of their network and make certain decisions to optimize the revenue by adjusting different parameters.