You are on page 1of 7

17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

GE 122 L6: INTRODUCTION


TO NUMERICAL METHODS

Ransie Apura
Assistant Professor
UPD DGE

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

▪ Definition
OUTLINE
▪ Numerics Covered in GE
122

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 2

1
17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

NUMERICAL METHODS (NUMERICS)


• Analytical methods provide exact solution as answer while numerical
methods provide table of values, graphs, etc.
• Can be a solution or approximation of solution to problems
• Choice of numeric method/algorithm best suitable for the problem and
how accurate results are.
• Estimation of errors, order of convergence, numeric methods in terms of
algorithms, etc.
• Complicated integral or roots of higher order polynomial or complex
differentiation.
• Systematic methods that uses simple algebra, graphical representations or
both.
• Iterative in nature (depends on a good starting value).

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 3

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

STEPS
1. Modelling – mathematical model of the problem.
2. Choose a numeric method and parameters (step size) –
preliminary error estimation.
3. Algorithm – programming.
4. Computation.
5. Interpretation of results – rerun if necessary.

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 4

2
17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

PROPER SKILLS FOR NUMERICS


• Programming
• Background on numerical methods
• Weighing trade-offs (e.g., simple and fast method but not
efficient with errors vs long method with small error in
convergence)
• Determining acceptable errors

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 5

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

ERRORS OF NUMERIC RESULTS


Method dependent (including its resolution). Numerics are mostly
approximations, hence, not exact and involves errors.

• Roundoff Error – rounding discarding all digits from some


decimal.
o Chopping is different from roundoff (not recommended.
Error can be larger than that in roundoff).
o Loss of significant digits (calculation has fewer correct digits
than the number from which it was obtained).
o Can be avoided by simple changes in algorithm.

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 6

3
17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

ERRORS OF NUMERIC RESULTS


• Experimental Error – measurement/experiment related.
Example: Computed Value vs Actual Value (Temperature)
Ideal Gas Law is given by PV=NRT 𝑃 = 1.00𝑎𝑡𝑚, 𝑉 = 0.100𝑚3
𝑁 = 0.00420 𝑚𝑜𝑙, 𝑅 = 0.08206
𝑃𝑉 1.00 0.100
Computing for 𝑇 = = = 17℃
𝑁𝑅 0.00420 0.08206
but the actual temperature measurement is 15℃, is the ideal gas law invalid?

Repeating the measurement with same R and N values while P is doubled


and volume is halved, actual temperature measurement is now 19℃.

How much more accurate our experimental results would need to be to


avoid this error?
UP Diliman Department of Geodetic Engineering
RJAA | GE 122: Mathematical Methods in Geodetic Engineering 7

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

ERRORS OF NUMERIC RESULTS


• Truncating Error – measurement/experiment related.
Example: Computed Value vs Actual Value (Temperature)
Ideal Gas Law is given by PV=NRT 𝑃 = 1.00𝑎𝑡𝑚, 𝑉 = 0.100𝑚3
𝑁 = 0.00420 𝑚𝑜𝑙, 𝑅 = 0.08206

• Truncating Errors – prematurely breaking off (e.g., Taylor Series


replaced by sum of its few terms).

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 8

4
17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

ERRORS OF NUMERIC RESULTS


"true" error = 𝑇𝑟𝑢𝑒 𝑉𝑎𝑙𝑢𝑒 − 𝐴𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛

"true" error
𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝐸𝑟𝑟𝑜𝑟 = 100%
𝑡𝑟𝑢𝑒 𝑣𝑎𝑙𝑢𝑒

Error Propagation – errors present at the beginning and later steps


propagate into computation and affect accuracy.

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 9

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

NUMERICS
Algorithms – procedure to solve or approximate a solution to the
problem or simply approximation procedures.
• Stable – algorithms with small changes in the initial data
produce correspondingly small changes in the final results.
• Unstable – not stable ☺
• Conditionally Stable – algorithms that are only stable for certain
choices of initial data. Stability depends on initial data.

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 10

10

5
17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

NUMERICS
Convergence – stopping criteria (numerics are iterative in nature).
• Rate of Convergence – some numerical methods converge
rapidly more than others. Rapid convergence usually require
more refined initial guesses and more complex programming
than a method with slower convergence.
• Selection of numerical method is subjective. Faster
convergence is not always a better option.
• Not all methods converge.

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 11

11

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

NUMERICAL METHODS IN GE 122


• Root Finding
• Interpolation
• Indirect Methods for Solving Linear Systems
• Numerical Differentiation
• Numerical Integration
• Numerical Solution to Ordinary Differential Equations

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 12

12

6
17/10/2021

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

REFERENCES
• Kreyszig, Erwin (2011). “Advanced Engineering Mathematics,
10th Edition.” John Wiley and Sons.
• Chapra, S. and Canale, R. (2010). “Numerical Methods for
Engineers, 6th Edition.”
• Burden, R. and Faires, J. (2011). “Numerical Analysis, 9th Edition.”

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 13

13

NOT FOR SHARING OR DISTRIBUTION


Only for the educational purpose of GE 122

Questions? ☺
raapura@up.edu.ph

UP Diliman Department of Geodetic Engineering


RJAA | GE 122: Mathematical Methods in Geodetic Engineering 14

14

You might also like