A Linear Decomposition of Index Generation Functions: Optimization Using Autocorrelation Functions
Tsutomu Sasao
This paper shows that autocorrelation functions are useful to find a decomposition of an index generation functions: F(x1, x2, . . . , xn ) = G(y1, y2, . . . , yp), where yj (j = 1, 2, . . . , p) are linear functions of x1, x2, . . . xn−1, and xn. It also shows a strategy to reduce the number of variables p to represent F(x1, x2, . . . , xn ) using an autocorrelation functions.