Efficient Resource Constrained Scheduling Using Parallel Two-Phase Branch-and-Bound Heuristics

Offloading Network Data Analytics Function to the Cloud with Minimum Cost and Maximum Utilization

“…In order to assess the effectiveness of our heuristic, Fig. 3 compares the average computation time for 600 experiments with two well-known algorithms: BB [14], and BC [15], for the three setups of VM pool sizes (small, medium and large). We can see that our MCMU heuristic is faster than BB and BC approaches, especially for large setups.…”

“…Authors in [14] proposed a Branch and Bound (BB) approach for resource constrained scheduling in two phases to reduce the computation time. Authors in [15] proposed a multilevel generalized assignment problem for minimizing the assignment cost of jobs to machines using Branch and Cut (BC).…”

“…We can see that MCMU91 dramatically decreases the average cost compared to the baseline (SP) and provides values that are close to the LP. The reason is that the baseline scheme selects the standard VM by default without considering the gNBs load and the matching VM capacity.In order to assess the effectiveness of our heuristic, Fig.3compares the average computation time for 600 experiments with two well-known algorithms: BB[14], and BC[15], for the three setups of VM pool sizes (small, medium and large). We can see that our MCMU heuristic is faster than BB and BC approaches, especially for large setups.…”