
When the iterative algorithm's speed depends on convergence, the direct algorithm's speed depends on its complexity. That's brilliant! Why do we need iterative algorithms at all? Also, since it has “an ultimate convergence” starting point selection doesn't matter. Since the algorithm is now direct, it doesn't need exit criteria. Here is the interactive illustration of it. Of course, it would be fantastic to start right at the solution point, but generally, our algorithm should work for any starting point we choose. Presumably, as we're getting closer, the leap distance should shorten, so at some point, we might consider it small enough to stop the operation.Īnd we can leave the question “where to start” unanswered. Therefore, by projecting a point from one line to another we will get closer and closer to the solution.Īs for exit criteria, we can simply measure how far we have to travel to make an iteration - a projective “leap” from one line to another. If you project an arbitrary point on a line, the projection will be closer to ane point of that line, including of course the desired solution.

The part of the algorithm that will bring us closer to the solution will be a simple projection.

A linear equation is a sum of weighted variables that equals a constant. Things as convergence, computational error, algorithmic complexity, - are all easy to show on a task as intuitive and graphic as solving a simple linear system.Ī linear system is a system of linear equations. This tutorial is all about the concepts that should help you recognize these problems and find the best solution for them.īut even if you are not interested in linear algebra at all, you might still find this tutorial interesting. Although, you will probably face the problems that may be solved in a form of linear systems, and often more elegantly than by any other means. This guide doesn't cover the implementation details of such, therefore. This is Words and Buttons Online - a collection of interactive #tutorials, #demos, and #quizzes about #mathematics, #algorithms and #programming.īeing a practicing programmer, you are very unlikely to implement yet another linear system solving algorithm all by yourself.
