摘要

BS (Binary Searching) is a known basic technique for keys searching in a sorted list. Fibonacci searching is another binary searching which splits remainder of the list by adopting no equilibrium criterion or irregular partition. For MFS(Minimal Fibonacci Searching), because each shift distance is always the shorter F(m-2) of F(m)=F(m-1)+ F(m-2) for i=1,2,...,k, if within f(m-1), shift distance is F(m-3) based on F(m-l)=F(m-2)+ F(m-3), and if within F(m-2), similarly, it is F(m-2)=F(m-3)+F(m-4), so that it's movement is shorter. This paper shows the irregular partition of Fibonacci and its PASCAL-likely algorithm, and also indicates, from the view point of searching length, Fibonacci is superior to the general binary searching, and also mainly in two ways that the algorithm of MFS is superior to that of BS, one just addition and subtraction to complete partition of sorted lists, and the other Min shift to search sorted lists. The latter is important if n is big enough and data is stored in file Storage.