Load balancing Optimization using Genetic Algorithm in comparison with Particle Swarm Optimization to handle Bag of Tasks in a Fog Environment

Abstract

Internet of Things (IoT) produces large-scale networks such as home appliances, high-tech
sensors, automated operations in Hospitals, Corporate Offices, etc., when the applications are
being analyzed which were probably considered for task computation and allocation of storages
based on resources utilized. Cloud environment initial levels were existing in databases and
resource utilization through the internet was not efficiently prioritized, which leads to slower
in daily usage for entire users. Such that Fog computing as an extension of the cloud
environment can be computed for IoT based resources. Devices such as smartphones, smart
watches that control IoT based appliances handle using an advanced architecture along with
Fog computing requirements. There are many research problems such as time elapse, more
complexity, high operational cost and much more. Data is transferred through many nodes from
various networks which are taking more time to distribute the IoT information from the
applications. In this situation resources are also calculated, and distribution based on
performance is taking more time and cost for resources. Therefore, the minimization is the key
point for reducing the latency and response time that can handle Bag-of-Tasks (BoT) for
scheduling multiple tasks using a fog environment. The proposed system is introduced with
Genetic Algorithm (GA) for optimizing the computed IoT data. Based on various levels of
schedulers for outperforming resources along with less tasks are compared with other
algorithms such as Round Robin Algorithm (RRA) and Particle Swarm Optimization
Algorithm (PSO). In the implemented results load balancing is highlighted based on all
optimization techniques and its performance which can schedule effectively using fog
environments. The obtained values from different dataset along with the size of tasks are
compared with Round Robin & PSO to achieve better response time for GA.