Optimal fog node selection based on hybrid particle swarm optimization and firefly algorithm in dynamic fog computing services

Abstract

Fog computing is a qualifiedly roseate technology introduced to support latency-sensitive and mission-critical applications by bringing resources closer to the end users. To exploit the full potentials of this auspicious technology, it is important to select optimal fog nodes for secured service provisioning. Therefore, in this work, an optimal fog node selection is formulated as a multi-objective optimization problem. To solve this problem, an efficient selection strategy based on improved particle swarm optimization (PSO) and modified firefly algorithm (FA) is developed. This study considers two important objectives that have not been optimized in the previous works: trust and rate of sojourn, in addition to remaining node capacity and energy consumption. The linear weighted-sum approach is used to aggregate the individual objective functions, after which the Best–worst method (BWM) is used to determine the weight vector of the aggregated function. The impact of the variation in the weight values of the most sensitive objective indicate that in the worst-case scenario, the proposed model is robust and insensitive with an improvement of about 42.16%–50.02% over the related methods. Moreover, we conduct an experimental simulation of the proposed selection strategy as well as comparative analysis with the state-of-the-art algorithms using six performance metrics. The results show that the proposed hybrid PSO-FA achieves a higher accuracy and faster convergence. The proposed solution also records an improvement of about 18.64%–69.45% in resource utilization, 10.94%–40.45% in energy consumption, 12.50%–75% in trust violation, 4.94%–31.12% in computation delay, and 24.71%–37.7% in makespan over the related advanced algorithms.