The virtual network functions (VNFs) placement problem has drawn significant attention from both academia and industry in recent years. Most of the researchers have ignored the fact that the probability of traffic flows through VNFs cannot always be 100%. In this paper, we study the placement scheme for virtual network function considering randomized data traffic (VNFPRAT). Our objective is to determine optimal deployment locations for VNFs and minimize total end-to-end delay. We formulate the VNFPRAT problem as a 0–1 nonlinear programming problem and prove its NP-hardness. This formulation is linearized to obtain the optimal solution for small scale networks. Besides, two efficient metaheuristics, i.e., greedy and simulated annealing, are proposed quickly find a near-optimal placement solution. Extensive simulations demonstrate that our proposed approach achieves 38.8% less end-to-end delay than the generic algorithm.

中文翻译：

---

随机到达组网流量下的网络功能布置

近年来，虚拟网络功能 (VNF) 放置问题引起了学术界和工业界的极大关注。大多数研究人员都忽略了一个事实，即流量通过 VNF 的概率并不总是 100%。在本文中，我们研究了考虑随机数据流量的虚拟网络功能的放置方案（VNFPRAT）。我们的目标是确定 VNF 的最佳部署位置并最大限度地减少端到端的总延迟。我们将 VNFPRAT 问题表述为 0-1 非线性规划问题并证明其 NP 硬度。该公式被线性化以获得小规模网络的最佳解决方案。此外，提出了两种有效的元启发式算法，即贪婪算法和模拟退火算法，可以快速找到接近最优的放置解决方案。