Читать книгу Algorithms For Dummies - John Paul Mueller, John Mueller Paul, Luca Massaron - Страница 41

Even when you find a good solution, one that is both efficient and effective, you still need to know precisely what the solution costs. You may find that the cost of using a particular solution is still too high, even when everything else is considered. Perhaps the answer comes almost, but not quite, on time or it uses too many computing resources. The search for a good solution involves creating an environment in which you can fully test the algorithm, the states it creates, the operators it uses to change those states, and the time required to derive a solution.

Often, you find that a heuristic approach, one that relies on self-discovery and produces sufficiently useful results (not necessarily optimal, but good enough) is the method you actually need to solve a problem. Getting the algorithm to perform some of the required work for you saves time and effort because you can create algorithms that see patterns better than humans do. Consequently, self-discovery is the process of allowing the algorithm to show you a potentially useful path to a solution (but you still have to count on human intuition and understanding to know whether the solution is the right one). The following sections describe techniques you can use to compute the cost of an algorithm using heuristics as a method of discovering the actual usefulness of any given solution.