Professional Documents
Culture Documents
6bb - Numerical Differentiation Up To 10th or Higher Order (08!11!2018) ... Edn 0
6bb - Numerical Differentiation Up To 10th or Higher Order (08!11!2018) ... Edn 0
To
Subject: Book – Numerical “derivative molecules (weight coefficients)” up to 10th or higher order
spatial derivatives
I have observed that national as well as international books on numerical differentiation have declared
differentiation molecules up to 6th derivative. The function has been expanded with Taylor’s series and
with Gauss Elimination Method, weight coefficients have been worked out. Then weight coefficients
have been used to solve partial differential equations.
The first principles have been employed, up to fourth derivative, to get coefficients. Further coefficients
can be predicted by induction method.
M
The the computation becomes bigger and bigger with higher order. manually, such computation may be
huge time consuming. This need triggered to explore new or better method with minimum computation
and less time consuming.
As a traditional or conventional method, 11 point, tenth order accuracy for tenth derivative has been
chosen to work out weight coefficients. The values can be compared and validated with classic paper of
W.G.Bickley, 1941. The key steps are:
For a known function, apply Taylor’s series expansion for steps e.g. f(x-5h), f(x-4h),f(x-3h),f(x-2h),f(x-h),
f(x),f(x+h),f(x+2h),f(x+3h),f(x+4h),f(x+5h). Now, assemble coefficients in a determinant of 11 by 11.
Apply Gauss elimination method to obtain 11 weight coefficients. Computation can be done in
MATHCAD-14 or later version or program code can be written. The book will cover at least 35 files; about 261 pages
Now, alternate method (induction method) can be employed. I am appending censored “applied
research report” as well as weight coefficients.
Interested publishers may come forward for full details. Prediction can be done for any order derivative.
Thanks.
N.T.Dadlani
nanak.dadlani@gmail.com
p.s
08-11-2018 (14-07-2017)
Applied Research:
Numerical Differentiation of Higher Orders & their application in PDEs
Table of Contents
Framework for research project (17-steps)
5) Augmentation of resources
6) Review of literature
7) Derive formulae from first principles
8) Formulation
9) Trend analysis
10) Prediction
11) Verification
12) Validation (including mathematical check)
13) Generalization (including mathematical check)
14) Evaluation of objectives
Submission of report:
1
1 off 19
08-11-2018 (14-07-2017)
After completing the research (in 2016) on “developing integration coefficients molecules” up to 10-dimensions
integration, it was felt that the method should also be tried for “numerical differentiation for higher orders”. I must
also explore the application of both molecules in various applications of engineering, science etc. through partial
differential equations of higher order say up to 10th order of spatial derivative etc. The ultimate aim is to make
mathematics simple & enjoyable.
CHAPTER-2
2. Research questions:
Is there any trend among coefficients of 4-functions for numerical differentiation as mentioned in table-1?
CHAPTER-3
3. Scope of research
3.1. Solve spatial derivates up to 10th order (second order truncation error with Taylor’s series expansion) through
analytical &/or numerical method. Establish link & develop “differentiation coefficients molecules” for higher
order PDEs. Ensure solution satisfies key objectives (#4).
3.2. Explore new or better method, through structured iterative research process.
Note:
The readers may differ with number as well as sequence of steps. You have to finally judge the outcomes.
“Conceptualization to generalization of process or up to the launch of the product” has to be properly articulated &
documented. Categorize method under: Incremental or breakthrough or radical innovation or invention?
3.3. Initially, explore derivation from first principles up to 4th order spatial derivative or pick up 4 functions, used for
decades, from any reputed book and then explore extension up to 10th order spatial derivative. I could do
both.
2
08-11-2018 (14-07-2017)
a.t b.x
- Develop a TRIAL FUNCTION e.g. u(x,t) = e .e
The trial function, ICs, & BCs must satisfy PDE. Find arbitrary constant a, b etc.
CHAPTER-4
3
08-11-2018 (14-07-2017)
Plan:
CHAPTER-5
5.2 Assess current & desired resources. Augment the resources as per requirements.
5.2.1 Engage masterminds / mentors / guide / domain experts in time to navigate & overcome the
bottlenecks in time.
5.2.2 MathCAD-14 & MATLAB 7.01 are available for expediting the calculations & assist in validation /
verification.
5.3.1 Ambition:
5.3.2 Passion to innovate, convert constraints into opportunity or challenge. Make mathematics simple and
interesting.
5.3.3 Contribute “unique knowledge” to society for solving complex problems
5.4 Knowledge-requirements:
5.5.1 Visualization (stretch imagination), analytical ability, people management, change management,
resources management etc.
5.5.2 Perseverance, exploration / experimentation, patience, accept failures
5.5.3 Project management techniques: Leadership etc.
5.5.4 Business acumen (entrepreneurship)
4
08-11-2018 (14-07-2017)
5.6.3.2 Commercial packages and their limitations (MathCAD-14, MATLAB7.01) for verification of past
examples and validation of new findings.
5.6.4 Research process: Conceptualization process (17-step
9 research process )
Standard examples &/or programs were identified from the books { }, & in articles { } referred in
bibliography.
Note:
Nobody (friend, assistant etc) understood my urge for doing this task, sense of urgency, etc. Why he /she would invest time for
me? A notional trial or attempt was made by others. It gave me some hint how to move further. I came to conclusion that I
have to do it myself from scratch.
CHAPTER-6
6. Review of literature
6.1. Analytical, & Numerical methods:
List out the status of current methods. Explore scope of innovation (explore new or better methods).
Identify & access top 5-articles as well as top-5 top books on the related topic. It should be done in consultation with
domain expert, product designer, professors, practicing engineers / scientists / mathematicians as the case may be. Collect
at least 5 master examples.
Review the literature for derivation of formulae from first principles, formulation, trend analysis, prediction,
validation, verification, generalization, & “meet technical-commercial-economic-social objectives & innovation
indices”. Here, Taylor’s series expansion has been considered.
It has been experienced that 80% time has been consumed in review of literature. Only 20% time is used for
execution or implementation for the given scope of work.
Note:
For product, additional steps are needed: Pilot run, feedback from peers / critics/ customers, refinement based
on feedback, re-testing of the product, scaling-up, & launch of the product.
4
Here three books are mentioned for employing various techniques / methods. Refer other books in bibliography.
# 1: A compendium of PDE models – Method of Lines (MOL) analysis with MATLAB, 2009
5
08-11-2018 (14-07-2017)
# 2: Travelling wave analysis of PDEs – Numerical and analytical methods with MATLAB & MAPLE, 2011
1. tanh method
2. exp method
3. Ricatti equation method
4. Direct integration method
5. Factorization method
6. Solutions by arbitrary constants
7. Other methods (MOL etc)
8. Maple built-in procedure for travelling wave equation
# 3:
Stanley J. Farlow
1. Separation of variables
2. Integral transforms
3. Change of coordinates
4. Transformation of the dependent variables
5. Numerical methods
6. Perturbation methods
7. Impulse-response technique
8. Integral equations
9. Calculus of variation methods
10. Eigen function expansion
Matthew N.O.Sadiku
Analytical techniques
1. Separation of variables
2. Series expansion
3. Conformal mapping
4. Integral solutions (Fourier transforms, Laplace transforms)
5. Perturbation methods
Numerical techniques
The answers to the research questions have to be probed here. Following is the starting point:
It was a challenge as well as opportunity either to derive numerical differentiation with first principles &/or
directly predict coefficients beyond 4th order spatial derivatives. Moreover, the accuracy of results, speed of
execution etc must be same or better but within permissible deviations.
The coefficients have been placed in table-1. Now, follow chapter-7 to chapter-14.
CHAPTER-7
fi+2(x) … f(x+2.h)
fi+1(x) … f(x+h)
fi … f(x)
fi-1(x) … f(x-h)
fi-2(x) … f(x-2.h)
========
When “exact function is unknown”, explore trial solution in Cartesian coordinates for odd order
or even order or mixed order (odd order + even order) spatial derivatives with exp or other method
• u(x,t) = e (a.t).e(b.x) … for all PDEs (1D PDE)
a.t
• u(x,t) = e( ). sin(b.x) …for even order spatial derivatives, 1D PDE
a.t
• u(x,t) = e( ).cos(b.x) …for odd order –do-, 1D PDE
• u(x.t) = e(a.t). {cos(b.x) + sin(b.x)}…for mixed order –do-, 1D PDE ?????
• u(x,y,t) = e(a.t).(ex).sin(y)…depending upon IC & BCs, 2D PDE
• u(x,y,z,t )= e(a.t).(ex).(ey).sin(20.5.z) .. –do- , 3D PDE
========
7.1 Derivation from first principles: At least four derivations must be done from first principles. Formulate 4-
functions or pick up 4-functions from the standard book as cited above. Then focus on to find trend or pattern or
behavior of coefficients. Refer table-1 for the coefficients, extracted from the 4-functions. Derive additional
formulae for 5th & 6th order derivative if possible. It will help for validation.
Here, Taylor series expansion has been used for first four derivations. Second order error truncation has been
applied. Further derivations for 5th order onwards are not required. Note following:
7.1.1 Ensure mathematical check as “sum of all coefficients must be zero for each row”.
7.1.2 Accurate prediction will avoid repeated and tedious derivations with first principles.
Note:
I have developed & successively applied above coefficients in numerical methods for differentiation & PDEs. Refer MathCAD file
dated 02-05-2017.
m
7.2 Methodology for creating “differentiation coefficients molecules”: (E - 1) , E=1, m=1,2,3,…,10
8
08-11-2018 (14-07-2017)
Refer: MathCAD (PDF file) for an example of PDE, which covers mixed order derivatives. The formulation is
slightly different for even-order with that of odd-order spatial derivatives.
(1 -1) =1 -1 …. (1)
How to “multiply” & then “add” coefficients: (Refer # 2): Create “differentiation molecule”
1 -1
1 -1
==========MULTIPLY
1 -1
x -1 1
=============================ADD
1 -2 1 …… (2)
===============================
If I hide first row or initial two rows, it would be extremely difficult to establish a trend by above method. I have
chosen all formulae or derivations up to fourth order, based on second order truncation error.
For higher accuracy of numerical results, I can choose different set of formulae with fourth order truncation error.
Refer appendix for table-4. I have yet to commence research whether or not the coefficients shown in table-4 are ok.
CHAPTER-8
8. Formulation
8.1. Place coefficients in proper sequence as shown in table-3, 3(A), & 3(B).
CHAPTER-9
9
08-11-2018 (14-07-2017)
9.1.2 Do we see any link or correlation in coefficients between two successive rows? If yes, then what is that
link? [1 -1]. When trend is visible? From 1st to 4th row.
9.1.3 Sum of all coefficients of each row is equal to zero.
9.1.4 By applying the “link“, predict coefficients of next higher order derivatives i.e. 5th & 6th order
derivative. Compare coefficients predicted for 5th & 6th order derivative with that of additional
derivations done for 5th & 6th order derivative. Is there congruency or any deviation? It is congruent &
no deviations are observed.
CHAPTER-10
10. Prediction
10.1. Predict coefficients (as mentioned in 9.1.4) for 5th to 10th or higher order. Develop “differentiation coefficients
molecule” for each spatial derivative.
10.2. Developing each molecule will take hardly 30 seconds (MANUALLY!).
CHAPTER-11
11. Verification
11.1. Verify the results of all past examples. Ensure key objectives have been met.
11.2. The “differentiation coefficient molecules” must be treated as unique because further calculation or solution
of PDE is being performed in “one step with pre-defined step h”. If required, step h can be increased or
decreased to get better results.
11.3. PDE is not converted to ODEs for obtaining the solution.
CHAPTER-12
12. Validation
12.1. Ensure mathematical check passes on each row (9.1.3) i.e. sum of all coefficients must be zero for each
row.
It must pass the validation test. Congruency must be visible or at the most deviation must be within tolerance when compared with
benchmark values (here, numerical values are compared with analytical values)
12.2. Numerical method with “differentiation molecules duly developed” gives solution with one iteration & a
pre-defined “step h” for each order of spatial derivative.
CHAPTER-13
13. Generalization
13.1. Pick-up any row & multiply with (1 -1) to obtain coefficients of next higher order derivative.
m
OR, Coefficients = (E - 1) , where E=1, m=1,2,3,….,10
13.2. Apply mathematical check as mentioned in prediction.
13.3. Now, write code.
10
08-11-2018 (14-07-2017)
CHAPTER-14
CHAPTER-15
Submission of report:
15. Summary of research (achievements, recommendations, limitations, enhancement for future): Accomplishment of
measurable & intangible targets:
15.1 Achievements: What has been achieved?
I was able to get first four formulae from extremely reliable book where derivations have already been done
and accepted by the domain experts as well as the required formulae have been used for decades.
Alternatively, I could derive 4 – formulae with first principles. I could do both.
• The next logical step was to find a “common link or correlation” between two rows of coefficients by
comparison, as explained in previous case. In all, 9-key technical steps were employed as mentioned in
research process namely:
o Review /Survey of literature, derivation from first principles, formulation, trend analysis, prediction,
validation, verification, generalization, meet key objectives (techno-commercial-economical -social
etc)
• Once we found a “common link” (1 -1), between two successive rows, for prediction. We can also say it is a
stage wise successive differentiation. Generalization has also been done. Also mathematical check was
applied on coefficients of each row i.e. sum of all coefficients must be zero and the check was found to be
accurate.
• “Differentiation coefficients molecule” beyond 4th order spatial derivatives could be achieved.
• Now, I am also exploring application of coefficients in solving higher order PDEs (partial differential
equations).
• Key objectives have been accomplished. 17 steps research process has been documented.
Vedic mathematics has only 16 sutras (formulae). Keen observation is the key to determine “correlation or
link”. Simple multiplication & addition were applied. It is a deduction method.
11
08-11-2018 (14-07-2017)
15.4 Scope for enhancement in future: What could be done for enhancement in future?
• Refer “limitations”.
• Explore “differentiation coefficients molecules” with Taylor’s series expansion with fourth order
truncation. Refer table-4. This table could be starting point.
• Explore whether “integration molecules” can be used for validation of the PDEs.
CHAPTER-16
16. Appendices
16.1. Table-1, 2-a, 2-b, & 3 (coefficients of molecules for Taylor’s series expansion with second degree
truncation)
16.2. Table-4 (coefficients of molecules for Taylor’s series expansion with fourth degree truncation)
16.3. Excel sheet showing examples from two books (17.8 & 17.9)
CHAPTER-17
17. Bibliography
17.1 The Numerical Method of Lines (MOL) – Integration of PDEs, 1991, W.E.Schiesser
17.2 Dynamic modeling of transport process systems, 1992, Silebi, & W.E.Schiesser
17.3 Recent development in numerical methods & software for ODE / DAE / PDE, 1992, G.D.Byrne, &
W.E.Schiesser (pp 97-123)
17.4 Computational mathematics in engineering and applied science, 1994, W.E.Schiesser
17.5 Computational transport phenomena, 1997, W.E.Schiesser
17.6 Adaptive method of lines (MOL), 2001, W.E.Schiesser (chapter-1)
17.7 Ordinary & PDE routines in …MATLAB, 2004, Lee, & W.E.Schiesser
17.8 A compendium of PDEs models: method of lines analysis with MATLAB, 2009, W.E.Schiesser,
and G.W.Griffiths
17.9 Travelling wave analysis of PDEs: numerical and analytical solutions with MATLAB & MAPLE,
2012, G.W.Griffiths & W.E.Schiesser
17.10 PDE analysis in biomedical engineering, 2013, W.E.Schiesser
17.11 Differential equations analysis in biomedical science & engineering: Differential equations with R,
2014, W.E.Schiesser
17.12 PDEs analysis in biomedical science & engineering: PDEs with R, 2014, W.E.Schiesser
17.13 An introductory global CO2 model, 2015, McHugh A.J, & W.E.Schiesser
17.14 Method of lines (MOL) PDE analysis in biomedical science & engineering, 2016, W.E.Schiesser (with
MATLAB programs)
17.15 Spline collocation methods for PDEs – With applications in R, 2017, W.E.Schiesser
17.16 Numerical methods for ODEs, II edition, 2008, J.C.Butcher (pp-146)
17.17 Fundamental numerical methods for electrical engineering, 2008, Stanislaw Rosloneic (ch-6)
17.18 Solving applied mathematical problems with MATLAB, Xue, & Chen
17.19 Handbook of non-linear PDEs, Polyanin
17.20 Numerical analysis using R, Graham W. Griffiths, 2016
17.21 The Numerical Methods in FORTRAN, 1990, M.G. Salvadori & McCormick
17.22 PDEs with Fourier series & BV problems, II edition, 2005, Nakhle’ H. Asmar (for Laplace equations in
2D & 3D PDEs)
17.23 Numerical Methods, Robert W. Hornbeck, 1982 (Chapter-3)
17.24 Numerical PDEs for Environmental Scientists & Engineers – a practical course, 2005 (Chapter-2)
17.25 Computational electromagnetic with MATLAB, fourth edition, 2018, Matthew N.O.Sadiku
12
08-11-2018 (14-07-2017)
Top-articles: # In RED
17.26 Formulae for numerical differentiation, W. G. Bickley, Mathematical Gazette, vol-25, issue 263,
pp 19-27,1941 (//www.cambridge.org/).
17.27 Generation of finite difference formulas on arbitrarily spaced grids, B.Fornberg, Mathematics of
computation, vol-51, no. 184, 1988, pp 699-706
17.28 B. Fornberg, 1992 (pp 97 – 124) in “recent developments of software…
17.29 Calculation of weights in finite difference formulas, B. Fornberg, SIAM Rev., vol-40, no.3, pp 685-
691, 1998
//http.lehigh.edu/~wes1/books/mol/library
Programs are available for some of the above books authored by W.E.Schiesser
13
08-11-2018 (14-07-2017)
Can we predict coefficients & fill up the table for V to X order derivatives? Any mathematical check on coefficients?
Note:
Assume you don’t have any knowledge of table-2.
===================================================================
Properties of functions for trend analysis: Table-2
Can we predict coefficients & fill up the table for I to IV order derivatives? Any mathematical check on coefficients?
Coefficients
I
II
III
IV
V 1 -5 10 0 -10 5 -1 - -
VI 1 -6 15 -20 15 -6 1 - -
VII 1 -7 21 -35 0 35 -21 7 -1 -
VIII 1 -8 28 -56 70 -56 28 -8 1 -
IX 1 -9 36 -84 126 0 -126 84 -36 9 -1
X 1 -10 45 -120 210 -252 210 -120 45 -10 1
Note:
14
08-11-2018 (14-07-2017)
I 1 -1 - - - - - - - - - 2
II 1 -2 1 - - - - - - - - 3
III 1 -3 3 -1 - - - - - - - 4
IV 1 -4 6 -4 1 - - - - - - 5
V 1 -5 10 -10 5 -1 - - - - - 6
VI 1 -6 15 -20 15 -6 1 - - - - 7
VII 1 -7 21 -35 35 -21 7 -1 - - - 8
VIII 1 -8 28 -56 70 -56 28 -8 1 - - 9
IX 1 -9 36 -84 126 -126 84 -36 9 -1 - 10
X 1 -10 45 -120 210 -252 210 -120 45 -10 1 11
Note: Properties of functions for trend analysis. Observe coefficients of the table-1:
Vedic method (Keen observation, Simple multiplication): Trend analysis: Refer point #
================================================================================
15
08-11-2018 (14-07-2017)
fi = f(x)
16
08-11-2018 (14-07-2017)
th
Table-4: Up to 4 order derivatives (fourth order truncation error)
Order of Coefficients for numerical differentiation points
derivative
fi+5 fi+4 fi+3 fi+2 fi+1 fi fI-1 fi-2 fi-3 fi-4 fi-5 -
I - -1 8 0 -8 1 - - - 4
II - - - -1 16 -30 16 -1 - - - 5
III - - -1 8 -13 0 13 -8 1 - - 6
IV - - -1 12 -39 56 -39 12 -1 - - 7
V
VI - - - - - - - - - - - -
VII -
VIII - - - - - - - - - - - -
IX
X - - - - - - - - - - - -
Refer:
1. Numerical PDEs for environmental scientists & engineers – A first practical course, 2005 (table 2.6),
Daniel R. Lynch
17
08-11-2018 (14-07-2017)
A compendium of PDE models - Method of lines analysis with MATLAB,
2009
Dr. W.E.Schiesser & Dr. G.W.Griffiths
Chapter-
2 1D PDE
Chapter-
10 2D PDE
Chapter- 3D PDE
18
08-11-2018 (14-07-2017)
Travelling wave analysis of PDEs - Numerical & Analytical methods with MATLAB & MAPLE,
2012
Dr. Graham W. Griffiths & Dr. W.E.Schiesser
19
APPENDIX-2 weight coefficients
wtrial02z3ptable1.m
1 off 5
10/26/17 12:15 AM MATLAB Command
coefficients Window
for m=4, 5 pts. 1 of 1
5 x 5 matrix
c =
m..> 0 1 2 3 4
0 0.0833 -0.0833 -0.5000 1.0000 alpha-1
0 -0.6667 1.3333 1.0000 -4.0000 alpha-2
1.0000 0 -2.5000 0 6.0000 alpha-3
0 0.6667 1.3333 -1.0000 -4.0000 alpha-4
0 -0.0833 -0.0833 0.5000 1.0000 alpha-5
wtrial04z5ptable1.m
10/25/17 11:49 PM MATLAB Command
Coefficients Window
for m=6, 7pts 7 x 7 matrix 1 of 1
cm=6
=
0 1 2 3 4 5 6 alpha-1
0 -0.0167 0.0111 0.1250 -0.1667 -0.5000 1.0000 alpha-2
0 0.1500 -0.1500 -1.0000 2.0000 2.0000 -6.0000 alpha-3
0 -0.7500 1.5000 1.6250 -6.5000 -2.5000 15.0000 alpha-4
1.0000 -0.0000 -2.7222 0 9.3333 0 -20.0000
alpha-5
0 0.7500 1.5000 -1.6250 -6.5000 2.5000 15.0000
0 -0.1500 -0.1500 1.0000 2.0000 -2.0000 -6.0000
alpha-6
0 0.0167 0.0111 -0.1250 -0.1667 0.5000 1.0000 alpha-7
wtrial06z7ptable1.m
10/26/17 12:32 AM MATLAB Command
Coefficients Window
for m=8, 9 pts. 9 x 9 matrix 1 of 1
c =
m...> 0 1 2 3 4 5 6 7 8
0 0.0036 -0.0018 -0.0292 0.0292 0.1667 -0.2500 -0.5000 1.0000 alpha1
0 -0.0381 0.0254 0.3000 -0.4000 -1.5000 3.0000 3.0000 -8.0000
0 0.2000 -0.2000 -1.4083 2.8167 4.3333 -13.0000 -7.0000 28.0000
0 -0.8000 1.6000 2.0333 -8.1333 -4.8333 29.0000 7.0000 -56.0000
1.0000 -0.0000 -2.8472 -0.0000 11.3750 -0.0000 -37.5000 0 70.0000
0 0.8000 1.6000 -2.0333 -8.1333 4.8333 29.0000 -7.0000 -56.0000
0 -0.2000 -0.2000 1.4083 2.8167 -4.3333 -13.0000 7.0000 28.0000
0 0.0381 0.0254 -0.3000 -0.4000 1.5000 3.0000 -3.0000 -8.0000
0 -0.0036 -0.0018 0.0292 0.0292 -0.1667 -0.2500 0.5000 1.0000 alpha9
m ..> 0 1 2 3 4 5 6 7 8 9 10
alpha1 1 0 -0.0008 0.0003 0.0068 -0.0054 -0.0451 0.0542 0.2083 -0.3333 -0.5000 1.0000 alpha1
2 0 0.0099 -0.0050 -0.0834 0.0834 0.5278 -0.7917 -2.1667 4.3333 4.0000 -10.0000
3 0 -0.0595 0.0397 0.4830 -0.6440 -2.7188 5.4375 8.6250 -23.0000 -13.5000 45.0000
4 0 0.2381 -0.2381 -1.7337 3.4675 6.5000 -19.5000 -17.0000 68.0000 24.0000 -120.0000
5 0 -0.8333 1.6667 2.3181 -9.2722 -6.7292 40.3750 15.7500 -126.0000 -21.0000 210.0000
6 1.0000 -0.0000 -2.9272 -0.0000 12.7417 -0.0000 -51.1500 0 154.0000 0 -252.0000
7 0 0.8333 1.6667 -2.3181 -9.2722 6.7292 40.3750 -15.7500 -126.0000 21.0000 210.0000
8 0 -0.2381 -0.2381 1.7337 3.4675 -6.5000 -19.5000 17.0000 68.0000 -24.0000 -120.0000
9 0 0.0595 0.0397 -0.4830 -0.6440 2.7188 5.4375 -8.6250 -23.0000 13.5000 45.0000
10 0 -0.0099 -0.0050 0.0834 0.0834 -0.5278 -0.7917 2.1667 4.3333 -4.0000 -10.0000
11 0 0.0008 0.0003 -0.0068 -0.0054 0.0451 0.0542 -0.2083 -0.3333 0.5000 1.0000 alpha11
12 m......> 0 1 2 3 4 5 6 7 8 9 10
wtrial10z11ptable1.m