Comparison of Swarm Algorithms For Fog Computing Schedulers

With the advent of increasingly smaller devices, computing has managed to permeate various locations, enabling automation and data extraction in different spheres of activity. We can apply Internet of Things (IoT) devices in these systems, giving rise to smart agriculture, smart cities, smart mobility, and more. Problems solved by IoT sometimes require connectivity and interoperability, as well as management of time-critical applications. Fog Computing networks are used to precisely meet these characteristics, no longer relying on the classical architecture of IoT. In this sense, scheduling of tasks in such systems usually is a NP-Complete problem, and consequently, it is not always possible to find the best solution for a particular problem. Heuristics have been proposed as schedulers, and in this work, we compare three of these proposals, namely the Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Artificial Bee Colony (ABC) algorithms. The comparison was performed using the iSPD simulator, and the results indicate that ABC presented the worst timing performance besides having slightly better efficiency in the use of hardware, while PSO had the lowest simulated times, as well as the best usage efficiency averages.