Heuristic Algorithm for Virtual Network Mapping Problem

Abstract

Nowadays, resource allocation for virtual networks (VNs) is brought as an imperative problem. For the characteristics of virtual networks, multiple virtual networks with different topo can co-exist on a shared infrastructure. A difficult point of problem is how to use the resource effectively and to satisfy the requirement of virtual network request. This problem is NP-hard. In this paper, we introduce a heuristic algorithm to solve this problem. To simulate a virtual network mapping problem in real world, we using two input data: an infrastructure network which is modeled by a connected graph and a set of virtual network request graphs with each graph contains their constraints, time and duration. The main purposes are to maximize the revenue and to minimize the cost when allocate the virtual networks to a substrate network. The experimental results are reported to show the efficiency of propose algorithm comparing to the Enhanced Greedy Node Mapping (EGNM) algorithm.