Discrete event simulation with 2D data

Question

I want to write a discrete event simulation. I downloaded bus traces (long. and lati. (x and y coordinates)) data from crowded group at Darmouth. When I run statistics tests, x and y data are correlated, so I cannot consider them as two separate distributions. Traditionally, we have a probability distribution, and we generate random numbers according to this distribution, this time I have 2D distribution. I am not clear about fitting how to fit this data to a distribution and generate random numbers to continue other steps in my DES.

Anyone have ideas? I would really appreciate.

Answer 1

It is not appropriate to think of this data in terms of a 2D distribution. There are spatial simulation techniques used to model data such as pollutant concentrations at a site, but buses run along specific routes. What you probably want to do is plot out the coordinates against the bus stop locations and model arrival times as a Poisson process.