Abstract:In this paper ,a new variational principle is proposed for the finite dimensional autonomous Hamiltonian systems and four types of symplecitc numerical algorithms are constructed .A modified ac-tion is defined by using the new variational principle .T hen ,the approximated action is obtained by ap-proximating the generalized coordinates and momenta by Lagrange polynomials within a time step ,and approximating the time integrals by means of Gaussian quadrature .Based on the approximated action and by taking generalized coordinates or momenta as independent variables at each end of the time step ,four types of symplecitc numerical algorithms are constructed .In this paper ,the detailed procedure for the construction of the algorithms is given and the proof of symplectic property and the numerical perform-ance of the algorithms will present in other papers .

The symplectic algorithms for Hamiltonian dynamic systems based on a new variational principle part I：the variational principle and the algorithms