Blind Search

Abstract

Full blind search assumes the exhaustion of all alternatives, where any previous search does not affect how next solutions are tested (left of Fig. 3.1). Given that the full search space is tested, the optimum solution is always found. Blind search is only applicable to discrete search spaces and it is easy to encode in two ways. First, by setting the full search space in a matrix and then sequentially testing each row (solution) of this matrix. Second, in a recursive way, by setting the search space as a tree, where each branch denotes a possible value for a given variable and all solutions appear at the leaves (at the same level). Examples of two quite known blind methods based on tree structures are depth-first and breadth-first algorithms. The former starts at the root of the tree and traverses through each branch as far as possible, before backtracking. The latter also starts at the root but searches on a level basis, searching first all succeeding nodes of the root and then the next succeeding nodes of the root succeeding nodes, and so on.