In this paper we constructed effective multivariate interpolation formulas for periodic functions, which are the precise on the Fourier polynomial classes. This paper continues investigations by N.M. Korobov [5], V.S. Rjaben'kii [11], S.M. Voronin [8], and others scientists on the application of the number-theoretic methods in numerical analysis. These authors was given the number of knots of a network equals to a prime number in the ring of integer rational numbers and in rings of integer numbers in algebraic numbers.
Here we consider the class of strictly regular periodic functions $f(x_1,\dots ,x_n),$ having the period on of one the each variables, and expanding in the absolute convergent Fourier series (see, for example, [15], p. 447) of the form
$$
f(x_1,\dots ,x_n)=\sum_{m_1=-\infty}^{\infty}\dots \sum_{m_n=-\infty}^{\infty}c(m_1,\dots ,m_n)e^{2\pi i(m_1x_1+\dots +m_nx_n)},
$$
where
$$
c(m_1,\dots ,m_n)=\int\limits_0^1\dots\int\limits_0^1f(x_1,\dots,x_n)e^{-2\pi i(m_1x_1+\dots +m_nx_n)}\;dx_1\dots dx_n.
$$
Further, we select the number of lattice points $N$ in the form $N=N_1\dots N_n,$ where $(N_s,N_t)=1$ as $s\ne t, 1\leq s,t\leq n,$ and $N_s\asymp N^{1/n}, 1\leq n,$ and using the Chinesse theorem on remainders, we construct the interpolation polynomial of the form
$$
P(x_1,\dots ,x_n)=\sum_{m_1=0}^{N_1-1}\dots\sum_{m_n=0}^{N_n-1}\tilde c(m_1,\dots ,m_n)e^{2\pi i(m_1x_1+\dots m_nx_n)},
$$
where
$$
c(m_1,\dots ,m_n)=\frac 1N\sum_{k_1=1}^{N_1}\dots \sum_{k_n=1}^{N_n}f\left(\frac{M_1^{*}k_1}{N_1},\dots ,\frac{M_n^{*}k_n}{N_n}\right)e^{-2\pi i\left(\frac{M_1^{*}m_1}{N_1}+\dots+\frac{M_n^{*}m_n}{N_n}\right)},
$$
moreover $N_sM_s=N, M_sM_s^{*}\equiv 1\pmod{N_s}.$
