A numerical approach to 2-D relativistic field theories is suggested. Considering a field theory model as an ultraviolet conformal field theory perturbed by a suitable relevant scalar operator one studies it in finite volume (on a circle). The perturbed Hamiltonian acts in the conformal field theory space of states and its matrix elements can be extracted from the conformal field theory. Truncation of the space at a reasonable level results in a finite dimensional problem for numerical analyses. The nonunitary field theory with the ultraviolet region controlled by the minimal conformal theory [Formula: see text] is studied in detail.

A program is proposed to study numerically the correlation functions in massive integrable 2D relativistic field theories. It relies crucially on the exact form factors of fields which can be reconstructed from the factorized scattering data. The correlation functions are expressible as infinite sums over intermediate asymptotic states. We suggest using computer power to perform the summation numerically. The convergence of the sum is tested for the simplest example of the scaling Ising spin-spin correlations (without magnetic field).

The critical Ising model perturbed by the spin field (conjugated to the magnetic field) is studied numerically by the method of truncated fermionic space of states. The matrix elements of the perturbed Hamiltonian are found between the states of free Majorana fermions living on a finite-length circle. The Hamiltonian spectrum is studied numerically in the level-5 truncated space. Reasonable estimations are obtained for the vacuum energy, a part of the mass spectrum and the simplest scattering amplitude of the perturbed system.

ABSTRACT

Using the vacuum correlator method the nonperturbative contribution to the P-wave spin-spin shift A = M , . , ( n ' p , ) -M ( n ' P I ) is calculated. It is shown that the sign and absolute value of A are very sensitive to the sign and amplitude u, of the nonconfining vacuum correlator D , . If u, < 0 then the shift A is also negative, but if n , > 0 we have obtained a positive A > 0 in contrast with the prediction in the framework of perturbative QCD.