Techniques for Developing Efficient Algorithms

Abstract

There are no general and standard techniques for designing efficient algorithms for a given problem. However, as you may have realised, we have been using certain useful techniques which have quite often helped us in designing efficient algorithms. For instance, binary search algorithms, tree operations and the quicksort algorithm all have a common theme: they divide the problem into two sub-problems of approximately equal size, and then recursively solve the two sub-problems. The best results are usually achieved when the problem is divided into two equal-sized sub-programs. Such a paradigm is known as divide and conquer and is discussed at length below. There are some other techniques which we shall discuss. It must be noted, however, that these techniques are merely guidelines and are not guaranteed to work for every problem. Nevertheless, when solving a new problem it is often useful first to try such a technique in the hope that it will give either a good solution or perhaps some insight regarding the problem properties.