Wavelet bases adapted to the settlement of infrared divergence (Q2566818): Difference between revisions

The author proposes a way to construct `flat' wavelets, which vanish to some order at the origin, by modifying a Daubechies wavelet basis \(\{\psi_{j,k}^\varepsilon, j\in{\mathbb Z}, k\in{\mathbb Z}^n, \varepsilon\in\{1, 2,\ldots,2^n-1\}\}\). The first step is to isolate, at a fixed scale \(j=0\), the wavelets which do not vanish at the origin or a derivative of which does not vanish. The second step is to impose cancellation at the origin on these wavelets by an explicit isomorphism in a Sobolev space. As the wavelets vanish to a sufficiently high order at the origin, the wavelet expansion is convergent in the distributional sense and form an unconditional basis for homogeneous Sobolev spaces \(\dot H^s({\mathbb R}^n)\). These bases also allow one to confine the infrared divergence (that is, the low frequency term) in the non critical cases \(s-n/2\not\in{\mathbb N}\). Such confinement can also be obtained by a wavelet basis with pseudo-polynomials, constructed from a Meyer adapted wavelet basis which is made of wavelets with support `away from 0' but some of them do not verify the conditions of vanishing moments. This method is applied to the Mumford process. In the critical case \(s-n/2\in{\mathbb N}\), a confinement of the infrared divergence is proposed in a different sense.