Control of a production-inventory system using a PID controller and demand prediction ⋆

A common and important problem in business is the determination of inventory policies for a production system within a changing business environment and market demand. In this paper, an automatic pipeline feedback order-based production control system (APIOBPCS), considering a demand with cyclic and stochastic components, is proposed. The dynamics and delays of the production process are modeled as a pure delay. The control system structure consists of a PID (Proportional, Integrative and Derivative) controller with an Extended Kalman Filter-based demand prediction. The main objective of the this dynamic controller is to stabilize and regulate the inventory levels in function of a desired set-point level. The Extended Kalman Filter (EKF) estimates the parameters of a Volterra time-series model to forecast future values of the demand. A control error analysis is also performed for the proposed inventory control system, in order to obtain bounds for the control errors and to probe its stability. This methodology is useful to make an appropriate decision about the desired inventory level for a given demand prediction error. The inventory control system is evaluated by simulations showing a good performance.