Levy walk is similar to random walk, but its walk time (flight time) and pause time are more complex. Each step is a tuple S=(l,q,Dtf,Dtp). In this model, each node takes a random direction q, and a flight length l. These values specify a destination point in or out the map. Dtf indicates the flight duration and picked for each flight from a probability distribution p(l). Using the Dtf and l, the model calculates speed of flight. Dtp specifies pause time at the end of a flight and it will be calculated using a configurable distribution y(Dtp).
In order to create random numbers for needed distributions, I used random generator algorithm presented in MATLAB codes that model creators uploaded in http://netsrv.csc.ncsu.edu/twiki/bin/view/Main/MobilityModelsDnld. The random generator needs some parameters which follow:
- minflight & maxflight: specifying the range of flight length.
- alpha
- beta
- minpausetime & maxpausetime: specifying the range of pause time at the end of flight.
- flightscale.
Figure 29: Traveling pattern of nodes in Levy Walk model
Figure 30: Spatial node distribution in Levy Walk model
|