One way to search for solution to a problem is simply to randomly or systematically generate a putative solution and then test whether it solves the problem. For example, given a game of Sudoku, one can generate squares with randomdigits in and then check if it satisfies the rules of Sudoku (e.g. each digit occurs exactly once in each row and column). This is usually a highlyinefficient way to search, as it ignores all of the structure in the {[search space}}. However, it can be used when the search space is small to exhaustively check all possibilities.
Generate and test can also be used as a startng point to geberate more compelx algi=orithms, for example, by generating partial solutions, or using heuristics to guide the generation process. It can aslo be used theoretically as a comparison point for other algorithms -- if your new complex algorithm is not performing better than generate and test, you know you need to return to the drawing board!
Defined on pages 57, 57, 57
Used on pages 57, 429, 557
Also known as generate and test
Generate and test -- finding solutions to the magic square