Some results about the cross-correlation function between two maximal linear sequences (Q1234646): Difference between revisions

Let \(\{a_j\}\) and \(\{a_{dj}\}\) be two maximal linear sequences of period \(p^n-1\). The cross-correlation function is defined by \[ C_d(t)=\sum_{j=0}^{p^n-2} \zeta^{a_{j-t}-a_{dj}} \qquad \text{for}\;t=0,1,\ldots,p^n-2 \] where \(\zeta=\exp(2\pi i/p)\). Finding the values and the number of occurrences for each value of \(C_d(t)\) is equivalent to finding the complete weight enumerator for the cyclic \((p^n-1,2n)\) code with parity-check polynomial which is the product of the recursion polynomials for the two maximal linear sequences. here properties of \(C_d(t)\) are investigated. An expression for \[ \sum_{t=0}^{p^n-2} C_d(t)C_d(t+\tau_1)\dots C_d(t+\tau_{n-1}) \] is derived. When \(\tau_1=\tau_2=\ldots=\tau_{n-1}=0\) this is an analogue to the Pless power moment identities which is often used in calculation of the Hamming weight enumerator. When \(d\not\equiv p^i\pmod{p^n-1}\) it is shown that \(C_d(t)\) has at least three different values. We also provide an upper bound on the number of different values of \(C_d(t)\) for some choices of \(d\). Further, the values and number of occurrences of each value of \(C_d(t)\) is determined completely for several new decimations \(d\) when \(C_d(t)\) has less than or equal to six different values. Numerical results and some conjectures are given.