Energy Minimization Partial Task Offloading With Joint Dynamic Voltage Scaling and Transmission Power Control in Fog Computing

In the fog network composed of dense terminal devices and fog servers, how to reduce the system energy consumption during task offloading is a challenging problem. To solve this problem, this article first formulates the energy consumption minimization problem of partial task offloading under delay constraints with dynamic voltage scaling (DVS) and transmission power control (TPC) techniques. Second, this problem was decomposed into two subproblems to solve: 1) the partial task offloading problem with optimal energy consumption under known matching between the terminal device and fog server and 2) the optimal matching problem between terminal devices and fog servers. For the first subproblem, the optimal solution is obtained through theoretical derivation, and the EOPCO-S algorithm is proposed to solve it. For the second subproblem, we transform the original problem into a weighted bipartite graph matching problem and propose the Kuhn–Munkres-based EOPCO-M algorithm to solve it. Finally, numerical simulations are carried out to verify the theoretical derivation and the effectiveness of the proposed algorithms. Experimental results show that the proposed algorithm can significantly reduce the energy consumption of fog networks compared with several baseline algorithms.