A novel niching genetic algorithm with heterosis for edge server placement

Nowadays, edge computing has been applied for providing various intelligent services, because it can provide ultra-low data transmission delay between user devices and service platforms, by placing several edge servers near users. Edge server placement (ESP) is one of the most critical problems, which greatly influences resource efficiency and service quality by deciding positions where each edge server is placed. As ESP problem is generally a multi-modal and discrete optimization problem with high dimension, this paper proposes a novel niching genetic algorithm (NichingGA) to provide ESP solutions with optimized average response time. The main idea of NichingGA is first dividing the population into several niches and establishing an elite niche. Then, inspired by heterosis in biology, NichingGA produces more diverse offspring by crossing individuals across niches, for improving the global search capability. NichingGA is evaluated by extensive experiments and proven to have better performance than 11 up-to-date approaches in improving the overall request response time.