Exploring The Trade-off between Efficient Service Placement Time and Optimized Fog Colonies Utilizing Advanced Genetic Algorithm with Diverse Network Topologies

Optimized Fog colony is one of the aid of fog computing, consisting of fog devices, enable efficient management of large fog domains. Well-designed fog colonies can operate independently, resulting in effectively resource utilization and improved system performance. However, there is a lack of optimization approaches for organizing fog devices into colonies and determining which fog colony should be used to stipulated and execute the services required by an IoT application is its primary challenge known as the “efficient service placement time” (ESPT). Contrary to the prevalent practice of treating ESPT as a single objective optimization problem, which often proves insufficient for accommodating the escalating complexities of engineering practice, our study takes a different approach. We present a modeling framework that considers the ESPT in fog computing as a constrained multi-objective optimization problem. Our secondary objective is to minimize the response time of services within the fog colonies. We employ the advanced elitist non-dominated sorting genetic algorithm (MS-NSGA) to optimize the fog colonies for constrained multi-objective service placement problem. We compare our experiment for three different network topologies with various configuration. The experimental result demonstrate the best trade-off between service placement time and proposed scheme for Barabasi-Albert network topology. Additionally, the results also indicate a clear trend towards reduced response time.