Cost-Efficient Deployment of Fog Computing Systems at Logistics Centers in Industry 4.0

Abstract

In Industry 4.0, the factories become increasingly smart and efficient through intelligent cyber-physical systems based on the deployment of Internet of Things (IoT), mobile devices, and cloud computing systems. In practice, the cloud computing system in a factory is managed in a centralized way, and hence may not afford heavy computing loads from thousands of IoT devices in the factory. An approach to address this issue is to deploy fog/edge computing resources nearby IoT devices in a distributed way to provide real-time computing responses on sites. This paper investigates deployment of an intelligent computing system consisting of a cloud center, gateways, fog devices, edge devices, and sensors attached to facilities in a logistics center. Except for locations of the cloud center and sensors that have been determined based on the factory layout, this paper establishes an integer programing model for deploying gateways, fog devices, edge devices in their respective potential sites, so that the total installation cost is minimized, under the constraints of maximal demand capacity, maximal latency time, coverage, and maximal capacity of devices. This paper further solves this NP-hard facility location problem by a metaheuristic algorithm that incorporates discrete monkey algorithm to search for good quality solutions and genetic algorithm to increase computational efficiency. Simulation verifies high performance of the proposed algorithm in deployment of intelligent computing systems in moderate-scale instances of intelligent logistics centers.