Parallel Genetic Algorithm to Extend the Lifespan of Internet of Things in 5G Networks
Ying Zhang, Weihong Yu, Xiaodong Chen, Jianhui Jiang · IEEE Access · 2020
With the development and popularization of 5G networks, the coverage problem of the Internet of Things (IoT) will encounter the massive-node problem. In this paper, we design a parallel genetic algorithm that divides the coverage problem of IoTs with massive nodes into many small problems and then solves these problems using Hadoop in parallel. First, the algorithm uses partitioning and grouping operations to degrade the scale of a large IoT and makes the coverage problem solvable. The algorithm then adopts the multi-objective programming-based genetic algorithm (MPGA) to solve the coverage problem. MPGA uses the fast non-dominated sorting to optimize the IoT coverage and node redundancy; it implements the preferential selection of non-critical nodes to maximize the length of the configuration sequence of working nodes. Finally, the parallel genetic algorithm uses uniform mutation and individual pruning to optimize the genetic algorithm internally and force its solving process to quickly converge toward feasible solutions. Experimental results confirm that the MPGA outperforms the existing algorithm on small IoTs in terms of coverage, the number of nodes, computing time, and the IoT lifespan. They also demonstrate that the parallel genetic algorithm successfully solves the coverage problem of IoTs with massive nodes and significantly extends the IoT lifespan.