Evaluation of Objective Function Descriptions And Optimization Methodologies For Task Allocation In A Dynamic Fog Environment

Abstract

Industry, healthcare, and various other sectors are rapidly adopting the Internet of Things to drive information and automation systems. However, as the number of devices increases, the number of information sent over the network increases as well, inducing network congestion and a potential latency increase. To ensure that demanding applications, such as smart vehicles, are supported in the current network infrastructure, we provide a general methodology of distributing software from the cloud toward the edge, reducing multiple objectives such as latency. In this research we define several problems in multi-objective distribution scenarios, and compare several methodologies for defining and solving the problem. Additionally, we propose a method for decreasing the problem complexity, improving performance with only slightly reduced accuracy.