|
|
|
|
|
Moving the Planets© Adam Hughes
In simulating the behavior of a real physical system, one of
the earliest steps is, by necessity, defining the model which
will be used. Usually, because of resource limitations, some
assumptions must be made. In the case of a system of planets
and other galactic bodies, we will make the assumption that
there is enough space between the entities that they will, for
the most part, see each other as single points on the horizon.
This is obviously a large simplification, but there are
other approximations that can be made later on, if need be,
to correct our picture of reality.
We can assign initial velocities and positions to our planets, methods of which we'll discuss at a later point. However, the real meat of the simulation comes when we start moving these objects around. Planets and suns and such are very large, definite objects. Because of this, their motions are dictated by the laws of classical mechanics, which makes our lives a little easier than if we had to consider quantum mechanics, too. Classical objects can be described using Sir Isaac Newton's equations of motion. For a particle residing at (x,y,z) with velocity (vx, vy, vz), Newton's equations tell us that dx/dt = (vx), dy/dt = (vy), dz/dt = (vz) and d(vx)/dt = - (Fx)/m, d(vy)/dt = - (Fy)/m, d(vz)/dt = - (Fz)/m, where a set of forces (Fx, Fy, Fz) has been introduced. This is a set of differential equations that we will attempt to solve in order to evolve the system from one time to another and another. The trick is that these equations are not exactly solvable, and so some approximate method must be used. Next time, we'll look at some of the algorithms used Go To Page: 1 2
The copyright of the article Moving the Planets in Scientific Computing is owned by Adam Hughes. Permission to republish Moving the Planets in print or online must be granted by the author in writing.
|
|
|
|