“…Some of the limitations of existing works are: (i) following the traditional hierarchical structure of fog-cloud, not making optimal use of the computing resources available in the fog layer and executing more services on cloud data centers, which increases the user response time [5,23,24,26,44], (ii) using hypothetical data instead of experimental data and, thus, uncertainty about the feasibility of implementing the algorithm in the real world [5,14,24,26,27,34,40,41,43,45], (iii) modeling the problem as an ILP and the complexity of problem solving when increasing the number of services and fog nodes on a large scale [21,22,32,44], (iv) ignoring priority for the applications and demands of services [5,14,18,[21][22][23][25][26][27]34,35,40,43], (v) ignoring mobility of fog nodes [3,14,18,[21][22][23]25,26,35,41,43,45], and (vi) improving only two or three QoS metrics…”