论文信息 - On average case errors in numerical analysis

On average case errors in numerical analysis

Abstract The error of a numerical method may be much smaller for most instances than for the worst case. Also, two numerical methods may have the same maximal error although one of them usually is much better than the other. Such statements can be made precise by concepts from average case analysis. We give some examples where such an average case analysis seems to be more sensible than a worst case analysis.

Erich Novak

[1] Henryk Wozniakowski,et al. Information, Uncertainty, Complexity , 1982 .

[2] Henryk Wozniakowski,et al. Average Case Optimality for Linear Problems , 1984, Theor. Comput. Sci..

[3] Henryk Wozniakowski,et al. A survey of information-based complexity , 1985, J. Complex..

[4] John Darzentas,et al. Problem Complexity and Method Efficiency in Optimization , 1983 .

[5] Grzegorz W. Wasilkowski. Some nonlinear problems are as easy as the approximation problem , 1984 .

[6] S. Zaremba. The Mathematical Basis of Monte Carlo and Quasi-Monte Carlo Methods , 1968 .

[7] S. Haber. A modified Monte-Carlo quadrature. II. , 1966 .

[8] Erich Novak,et al. Eingeschränkte Monte Carlo-Verfahren zur Numerischen Integration , 1985 .

[9] Henryk Wozniakowski,et al. A general theory of optimal algorithms , 1980, ACM monograph series.

[10] Some approximate integration formulas of statistical interest , 1963 .

[11] Eric Novak,et al. Quadrature and widths , 1986 .

[12] Grzegorz W. Wasilkowski,et al. Average case optimality , 1985, J. Complex..

[13] H. Woźniakowski,et al. Can adaption help on the average? , 1984 .

[14] F. M. Larkin. Gaussian measure in Hilbert space and applications in numerical analysis , 1972 .