Supervisory control and data collection policies for a distribution center modeled as a discrete event system

Abstract

A large scale discrete event model of a distribution center is presented where critical parameters identified were utilized to create a real-time scheduling control policy to improve throughput and redistribute work in process (WIP), reducing bottleneck process and overall WIP. In implementing the control policy, it was found that the data was queried multiple times at the same time stamp and continuously throughout the process during a time when the control policy specification would not be met. The amount of data collection, processing and storage can become expensive especially when implementing multiple control policies that are dependent on several critical parameters. There is a need for smart data collection, processing and storage policies for cost effectiveness and to manage existing power management and battery life issues that may occur with continuous data streaming with existing wireless devices and sensors. This paper presents both the real-time scheduling control policy and the data collection policy for a distribution center that must be considered simultaneously in order to realize the maximum benefits of a wireless real-time scheduling system.