Computer modelling is not difficult, but it requires to know some calculus and some programming language. Most laws of physics become differential equations when it comes to calculate motion or flow of something. Say, if your task is to model comet's orbit around Sun - then you take Newton law of gravity: F=-GMm/r2 and write a differential equation: m(d2t/dt2)=-GMm/r2). But unlike in this simple case, many differential equations do not have analytical solution, thus have to be solved numerically. (Also, if there is one more gravitating body around your asteroid or on its way, then there is no analytical solution for such gravitating system too.) Numerical differentiation (and integration) is to split your process into small time intervals - say, from ti to ti+1 and replace derivatives within this interval by ratios of finite differencies like: dr/dt = (ri+1-ri)/(ti+1-ti) and d^2r/dt^2=(ri+1 -2ri+ri-1)/((ti+1-ti)(ti-ti-1), thus instead of differential equation you have a simple algebraic equation. Computer can solve it in a millisecond or so (using C, C++ or fortran), so you should write a simple code which starts from given initial or boundary values of r and dr/dt and determines ri, and then repeats it step by step using previous values to calculate next ones, thus showing, for example,dot-to-dot orbit of asteroid around mass M or perturbation of its orbit when passing by Earth. As r is a vector in 3-D, it has 3 components, so you actually should have a set of 3 algebraic equations for x,y, and z components of r.

So you have to learn some programming on some language to begin computer modelling. There are textbooks on numerical computation methods and practically all colleges today teach classes on numerical modelling on their computer science departments.

Also, when you have multi-body system or complex shapes/processes involved, you will soon need faster than your PC computer - but this is your professor's headache, not yours. He will buy some time on supercomputers, which usually have fortran, C++ and other common compilers as well as internet connection, so you can upload your code and dounload results right from/to your PC.

Results usually are in form of long colomns of 8-digit numbers in .31415926E+12 format (say, x,y,z components of r, and time t), and it is up to you to create another code to plot the orbit on paper (for publication) or dot-to-dot animate it on the computer display (for professor's presentation on coming conferences).

All equations of physics do not go beyond 2-nd derivatives of space and time, so main problem usually is not physics (wich actually is not even your problem - your professor will write all equations for you) but to keep track of coordinates and i, i-1, i+1 indexes in a set of computer equations and in multiple "if-then" loops inside loops inside loops of your code.

Fortunately almost any error immediately shows up in wierd numbers you get out of calculations - and you always can test your code by setting initial parameters to get some well-known solution or approximation, (say, for M=0 it should be a straight line, or for dr/dt=orbital velocity it should be a circle, etc) so developing code is not challenging problem.

And at the same time you learn physics, play computer games or chat internet when porofessor is not around (all grad students do), discuss origin of life (or of universe, or fields, or of God(s)) with your professor and other grad students around, travel from time to time to conferences or telescopes in exotic destinations, to other universities or research center to discuss you results or get someone's data or help, work part-time, date someone, etc.