Random matrix

from Wikipedia, the free encyclopedia

In probability theory and statistics (with main applications in theoretical physics ) a random matrix is a matrix-valued random variable ( English Random Matrix ). To distinguish it from the multivariate distributions, the distribution of a random matrix is called a matrix-variable probability distribution .

Many important properties of physical systems can be formulated mathematically with matrices and random matrices appear in problems of statistical mechanics. For example, the thermal conductivity of a crystalline solid can be calculated directly from the so-called dynamic matrix of the particle-particle interaction of the crystal lattice.

To motivate: Disordered systems

In the case of a disordered physical system (e.g. with so-called amorphous material) the relevant matrix elements are random variables. The physics of these systems can essentially be determined by the parameters of the respective matrices, e.g. B. by mean value and fluctuation of the respective size. The eigenvectors and the eigenvalues ​​of the random matrices are of special interest .

Spectral theory of random matrices

Mathematicians and physicists have developed many remarkable theoretical relationships and empirical evidence on the theory of random matrices. One of the main results is the so-called Wigner'sche law (see Eugene Wigner ): It states that the spectral measure of the eigenvalues of a symmetric random matrix, in physics known as the so-called density , a characteristic semicircle distribution is sufficient. It is about N  ×  N matrices with Gaussian elements in the Limes . Wigner's law applies not only to symmetric matrices (so-called orthogonal ensemble), but with slight modifications also to unitary or symplectic matrices (so-called unitary or symplectic ensemble).

Applications

literature

Web links

References and comments

  1. arxiv : math.OA / 0412545
  2. arxiv : math.OA / 0501238
  3. VS Rychkov, S Borlenghi, H Jaffres, A Fert, X Waintal: Spin Torque and Waviness in Magnetic Multilayers: A Bridge Between Valet-Fert Theory and Quantum Approaches . In: Phys. Rev. Lett. . 103, No. 6, August 2009, p. 066602. doi : 10.1103 / PhysRevLett.103.066602 . PMID 19792592 .
  4. DJE Callaway: Random Matrices, Fractional Statistics, and the Quantum Hall Effect . In: Phys. Rev., B Condens. Matter . 43, No. 10, April 1991, pp. 8641-8643. doi : 10.1103 / PhysRevB.43.8641 . PMID 9996505 .
  5. M Janssen, K Pracz: Correlated random band matrices: Localization-delocalization transitions . In: Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics . 61, No. 6 Pt A, June 2000, pp. 6278-86. doi : 10.1103 / PhysRevE.61.6278 . PMID 11088301 .
  6. DM Zumbühl, JB Miller, CM Marcus, K Campman, AC Gossard: Spin-orbit Coupling, Antilocalization, and Parallel Magnetic Fields in Quantum Dots . In: Phys. Rev. Lett. . 89, No. 27, December 2002, p. 276803. doi : 10.1103 / PhysRevLett.89.276803 . PMID 12513231 .
  7. ^ SR Bahcall: Random Matrix Model for Superconductors in a Magnetic Field . In: Phys. Rev. Lett. . 77, No. 26, December 1996, pp. 5276-5279. doi : 10.1103 / PhysRevLett.77.5276 . PMID 10062760 .
  8. Antonia M. Tulino, Sergio Verdu: Random Matrix Theory and Wireless Communications . Now , 2004.
  9. ^ D Sánchez, M Büttiker: Magnetic-field Asymmetry of Nonlinear Mesoscopic Transport . In: Phys. Rev. Lett. . 93, No. 10, September 2004, p. 106802. doi : 10.1103 / PhysRevLett.93.106802 . PMID 15447435 .
  10. F Franchini, VE Kravtsov: Horizon in Random Matrix Theory, the Hawking radiation, and Flow of Cold Atoms . In: Phys. Rev. Lett. . 103, No. 16, October 2009, p. 166401. doi : 10.1103 / PhysRevLett.103.166401 . PMID 19905710 .
  11. newscientist.com