BENNS is a benchmarking and Neural Network based surrogate model for approximating traffic latency across SFCs in and SFC embedding. This surrogate model is used in Genetic Algorithms to speed up the evaluation of candidate solutions. BENNS benchmarks SFC embeddings to encode them to two numerical values, which are then used to train a Neural Network to predict the traffic latency of the SFC embeddings. BENNS reduces evolution time by 98% while achieving near-optimal solutions.
Publications
BENNS: A Surrogate Model for Hybrid Online-Offline Evolution of SFC Embedding