An author says that Poisson process model may under estimate the incoming traffic especially when the traffic is heavy-tailed. He advises me to use Pareto distribution to model the traffic in cognitive radio networks. However, I observe that Poisson process has largely been utilized in the literature for the incoming traffic modeling in CRNs. Suggestions please?
Dear Abdullah
as you mentioned, Poisson distribution has been suggested for the arrival of vehicles in literature.
If you have the arrival data, you can use some statistical software (such as EasyFit) to determine the data distribution.
If you have data, you can try both distributions (if not more) and see which one is better for your application. Distributions are characterized by some parameters (shape and scale parameters) that you can find if you have a good data set. You can use some type of statistical test (for example Chi-square Test of Goodness of Fit) to determine which distribution fits better your data.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators