Joint optimization of energy consumption and time delay in IoT-fog-cloud computing environments using NSGA-II metaheuristic algorithm

• Published in: Journal of ambient intelligence and humanized computing Vol. 14; no. 3; pp. 1675 - 1698
• Main Authors: Jafari, Vahid, Rezvani, Mohammad Hossein
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-03-2023; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence; Augmented reality; Cloud computing; Computation offloading; Computational Intelligence; Edge computing; Energy consumption; Engineering; Evolutionary computation; Genetic algorithms; Heuristic; Heuristic methods; Internet of Things; Methods; Minimax technique; Network latency; Nodes; Optimization; Original Research; Queuing theory; Response time; Robotics and Automation; Search algorithms; Sorting algorithms; Statistical analysis; Time lag; User Interfaces and Human Computer Interaction; Virtual reality; Fog computing; Time-constraint task; Energy consumption; Internet of Things (IoT); Genetic algorithm; Task offloading
• ISSN: 1868-5137; 1868-5145
• DOI: 10.1007/s12652-021-03388-2

Today, there exists a growing demand for Internet of Things (IoT) services in the form of vehicle networks, smart cities, augmented reality, virtual reality, positioning systems, and so on. Due to the considerable distance between the IoT devices and the central cloud, using this option may no longer be a suitable solution for delay-constraint tasks. To overcome these drawbacks, a complementary solution called fog computing, also known as the
cloud at the edge
is used. In this solution, nodes at the edge of the network provide resources for IoT applications. Although offloading tasks on the fog nodes save energy on IoT devices, it increases task response time. Therefore, making a trade-off between energy consumption and latency is crucial for IoT devices. Because offloading falls into the category of NP-hard knapsack problems, metaheuristic methods have been widely used in recent years. In this paper, we formulate the problem of joint optimization of energy consumption and latency in the form of a multi-objective problem and solve it using the non-dominant sorting genetic algorithm (NSGA-II) and Bees algorithm (BA). Also, to improve the quality of solutions, we combine each of these methods with a robust type of differential evolution approach called minimax differential evolution (MMDE). This combination moves the solutions to better areas and increases the convergence speed. The simulation results show that NSGA-based methods have remarkable robustness compared to BA-based methods in terms of significant criteria such as energy consumption, time delay, and so on. Our statistical analysis shows that both NSGA-based and BA-based metaheuristic methods not only do not significantly increase energy consumption but also drastically reduce response time.