Computation of the multivariate normal integral

ABSTRACT

This paper presents a direct computation of the multivariate normal integral by the Gauss Quadrature method. An error control method is given. Results are presented for multivariate integrals consisting of up to twelve normal distributions. A computer program in FORTRAN is given.

REFERENCES

• 1: ANKLESARIA, K. P., AND DREZNER, g. A multivariate approach to estimating the completmn time for PERT networks. J Opel'. Res. Soc. 37 ~1986), 811-815.
• 2: DONGARRA, J. J., MOLER, C. B., BUNCH, J. R., AND STEWART, G.W. LINPACK Users' Grade. SIAM Publications, Ph}adeIphia, Pa., 1979.
• 3: DREZNER, Z. Computation of the bivariate normal integral. Math. Cornput. 32 {1978), 277 279.
• 4: DREZNI~'R, Z., ,a24D WESOLOWSKY, G.O. On the computation of the bivariate normal integral. J. Star. Comput. Szmul. 35 (1990), 101 107.
• 5: JOHNSON, N. L.~ AND KOTZ, S. Distributions in Statistzcs: Continuous Mult~vartate Dzstrzbutions, Wiley, New York, 1972.
• 6: SCHERVlSH, M. J. Multivariate normal probabilities with error bound. Appl. Star. 33, (1984), 81 87
• 7: SCHERVISH, M. J Correction to Algorithm AS 195: Multivar/ate normal probabilities with error bound. Appl. Stat. 34 (1985), 103-104.
• 8: STEEN, N. M., BYRNE, G. O., AND C*ELBARD, E. M Gaussian quadratures. Math. Comput 23 (1969), 661 671.
• 9: Su~c}{, R., AND TUREK, R. An asymptotic test for the Goodman-Kruskal A. Forthcoming.

INDEX TERMS

Primary Classification:
  G. Mathematics of Computing
  G.3 PROBABILITY AND STATISTICS
      Subjects: Statistical software

Additional Classification:
  G. Mathematics of Computing
  G.1 NUMERICAL ANALYSIS
      G.1.4 Quadrature and Numerical Differentiation
          Subjects: Multidimensional (multiple) quadrature

General Terms:
Algorithms, Performance

Keywords:
multivariate normal probability