Ant Mobility Model Platform for Network Simulator

摘要

Various ad hoc routing algorithms were proposed to increase the performance in ad hoc network. In order to estimate the performance of routing algorithms, they always simulate in ad hoc network environment, and a mobility model is also used for network simulation. There are two types of mobility models: trace and synthetic models. Trace models are more accurate than synthetic ones, and their simulation results are more accurate, too. However, new routing network environments are not easily modeled if traces have not yet been created. That is why most of the simulations are based on synthetic models. In this paper, a tool is implemented for generating a special trace model, called ant mobility model. It mimics the ants' movements of an ant colony. The movement is similar to the workers in a store. We focus on the integration between the ant mobility model and the network environment. When users setup the parameters for generating an ant mobility model, the generated model can be used directly in the network environment In order to illustrate the feasibility of ant mobility model, it is compared with the random waypoint mobility model based on the same routing algorithm, AODV, by using NS-2 simulator. Three major performance metrics, throughput, network latency, and control overhead message, are estimated. As we expected, these three metrics of the ant mobility model are better than that of the random waypoint mobility model. We suggest that a trace model, like ant mobility model, should be integrated quickly with the network environment in the same way.