Computational methods

This page describes computational methods which can be used to solve problems.

• Backtracking
  • Backtracking is an algorithm (usually recursive) that methodically checks every path and builds a solution based on the correct paths.
  • If an invalid path is found, the algorithm backtracks to the previous stage and visits an alternate path.
  • An example of backtracking is the depth-first search algorithm used for searching binary trees.
• Data mining
  • Data mining is a technique to spot patterns and correlation in large sets of data.
  • By spotting patterns and trends, predictions about the future can be made.
• Heuristics
  • Heuristics are educated guesses used when it is unfeasible to compute all eventualities.
  • By using an educated guess, the problem can be solved with a "good enough" result, although it is not 100% accurate.
  • Heuristics are used in the A* algorithm. An estimate for the distance to each node is given. This speeds up the process of finding a fast route to the end point.
• Performance modelling
  • Performance modeling involves testing a program using mathematical models with different operating systems.
  • This method is used to test safety-critical computer systems, for example, programs used to fly planes. Performance modelling can be used to test the capabilities of a system without the risks involved with real world testing.
• Pipelining
  • Pipelining is a process used to develop projects. The program is split into modules and which are developed in parallel with separate teams.
  • Typically, the output of one process becomes the input of another process, similar to a production line.
• Visualisation
  • Visualisation is the presentation of data in a way which is easier to understand
  • Examples are graphs, trees, charts, tables and matrices.