It is often the case in computer science (and in life) that we need to solve some particularproblem, and there isn’t just one unique answer that works as a valid solution. There could be multipleanswers that are all acceptable or close enough. For example, real world solutions to problems liketurning the dial to tune in a radio station on shortwave, pointing an antenna for best reception, oradding sweetener to your morning coffee don’t require an exact solution, just one that is “goodenough.” If you can hear the station clearly, the dial and antenna are sufficiently correct. For themathematical versions of this class of problems, there are frequently formulas and approximationsavailable that will give an answer that is “good enough” by using techniques such as successiveapproximations, least mean squares, stepwise refinement, or gradient descent.
展开▼
