Computation Offloading and Task Scheduling Based on Improved Integer Particle Swarm Optimization in Fog Computing

Abstract

As an extension of cloud computing, fog computing is a promising computing paradigm. By placing resources near terminals, users can access the fog servers with lower latency than cloud and utilize server resources to complete offloading tasks for optimizing QoE. Task scheduling in fog computing is to map requested tasks to appropriate fog node, so as to improve resource utilization efficiency and drop the cost of IoT devices. In this paper, we focused on the scheduling problem of offloaded tasks in fog scenarios with multiple users. We first formulated the problem as a combinatorial optimization problem, and proposed an improved integer particle swarm optimization method to solve the problem in a reasonable time. Experimental results showed that the proposed algorithm can reduce the running time by 90% compared with the traversal method, while resulting an approximate optimal value. It has good performance in the optimization of task scheduling problems.