Des Bornes Exactes De La Complexite Des Algorithmes Sequentiels De Recherche D’un Motif

String matching is the problem of finding all the occurrences of a string, called the pattern, within another, called the text. We give sharp bounds when only one head moving from the left to the right of the text is used, for the general case where the alphabet is unknown to the algorithms : b(2 − 1 m) nc comparisons between pattern characters and text characters, and min{1 + blog2mc, card(A)} comparisons on each text character, where m is pattern length, n is text length, and A is the alphabet. These results improve the previous results given by Knuth, Morris and Pratt, and, more recently, by Simon.