The application of symbolic mathematics to a singular perturbation problem

ABSTRACT

A basic technique for the numerical solution of ordinary differential equations is to express them as a singular perturbation problem. However, computational studies indicate that the resultant matrix equations which must be solved are often highly ill-conditioned. In this paper a particular singular perturbation problem which was shown to be ill-conditioned using 8 numerical methods is solved by symbolic techniques. These techniques lead both to an analytic proof of the solution plus to the precise knowledge of the asymptotic behavior of the solution vector as it converges. The difficulties encountered in solving the problem symbolically are discussed. Then several conclusions are drawn about the merits of a symbolic versus a numeric approach when applied to the solution of linear systems. Finally some advice and warnings to both the user and the designer of symbol manipulation systems are given concerning their goals and expectations when large matrix equations are to be solved.

REFERENCES

• 1: George E. Collins, The SAC-1 system: An introduction and survey, Proceedings of the second ACM symposium on Symbolic and algebraic manipulation, p.144-152, March 23-25, 1971, Los Angeles, California, United States [doi>10.1145/800204.806279]
• 2: Dorr, F.W., "The Numerical Solution of Singular Perturbations of Boundary Value Problems", SIAM Journal on Numerical Analysis, Vol. 7, No. 2, June 1970, pp. 281-313.
• 3: Dorr, F.W., "An Example of Ill-Conditioning in the Numerical Solution of Singular Perturbation Problems", Mathematics of Computation, Vol. 25, No. 114, April 1971, pp. 271-283.
• 4: Horowitz, E., "User Notes for the SAC-1 system at Cornell University", Computer Science Department, Ithaca, New York, 1970.
• 5: Donald E. Knuth, The art of computer programming, volume 1 (3rd ed.): fundamental algorithms, Addison Wesley Longman Publishing Co., Inc., Redwood City, CA, 1997
• 6: McClellan, M.T., The Exact Solution of Systems of Linear Equations with Polynomial Coefficients, Ph.D. Thesis, University of Wisconsin, Sept. 1971.
• 7: O'Malley Jr., R.E., "Topics in Singular Perturbations", Advances in Mathematics, Vol. 2, 1968, pp. 364-470.
• 8: Wasow, W., Asymptotic Expansions for Ordinary Differential Equations, Interscience, New York, 1965.
• 9: Shalhav Zohar, Toeplitz Matrix Inversion: The Algorithm of W. F. Trench, Journal of the ACM (JACM), v.16 n.4, p.592-601, Oct. 1969 [doi>10.1145/321541.321549]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.1 SYMBOLIC AND ALGEBRAIC MANIPULATION
      I.1.0 General

Additional Classification:
  G. Mathematics of Computing
  G.1 NUMERICAL ANALYSIS
      G.1.7 Ordinary Differential Equations

General Terms:
Theory

Keywords:
Exact solution of linear systems, Navier-stokes equations, Singular perturbation problem, Symbol manipulation