Example 9.  Plot the absolute error  [Graphics:Images/NumericalDiffMod_gr_260.gif]  over the interval  [Graphics:Images/NumericalDiffMod_gr_261.gif],  and estimate the maximum absolute error over the interval.  
9 (a).  Compute the error bound  [Graphics:Images/NumericalDiffMod_gr_262.gif]  and observe that  [Graphics:Images/NumericalDiffMod_gr_263.gif]  over  [Graphics:Images/NumericalDiffMod_gr_264.gif].  
9 (b).  Since we the function f[x] and its derivative is well known, and we have the graph for [Graphics:Images/NumericalDiffMod_gr_265.gif],  we can observe that the maximum error on the given interval occurs at x=0.  Thus we can do better that "theory", we see that [Graphics:Images/NumericalDiffMod_gr_266.gif]  over  [Graphics:Images/NumericalDiffMod_gr_267.gif].

Solution 9.

9 (a).  Compute the error bound  [Graphics:../Images/NumericalDiffMod_gr_268.gif]  and observe that  [Graphics:../Images/NumericalDiffMod_gr_269.gif]  over  [Graphics:../Images/NumericalDiffMod_gr_270.gif].  

[Graphics:../Images/NumericalDiffMod_gr_271.gif]

[Graphics:../Images/NumericalDiffMod_gr_272.gif]

[Graphics:../Images/NumericalDiffMod_gr_273.gif]

[Graphics:../Images/NumericalDiffMod_gr_274.gif]

9 (b).  Since we the function f[x] and its derivative is well known, and we have the graph for [Graphics:../Images/NumericalDiffMod_gr_275.gif],  we can observe that the maximum error on the given interval occurs at  [Graphics:../Images/NumericalDiffMod_gr_276.gif].  Usually we can do better that "theory", with [Graphics:../Images/NumericalDiffMod_gr_277.gif]  over  [Graphics:../Images/NumericalDiffMod_gr_278.gif]. However, the two bounds are nearly the same for this example.

[Graphics:../Images/NumericalDiffMod_gr_279.gif]

[Graphics:../Images/NumericalDiffMod_gr_280.gif]


[Graphics:../Images/NumericalDiffMod_gr_281.gif]

[Graphics:../Images/NumericalDiffMod_gr_282.gif]


[Graphics:../Images/NumericalDiffMod_gr_283.gif]

[Graphics:../Images/NumericalDiffMod_gr_284.gif]

[Graphics:../Images/NumericalDiffMod_gr_285.gif]

[Graphics:../Images/NumericalDiffMod_gr_286.gif]

Remark. Is there a difference between the error bound and minimal error bound?  It looks different i the 10 decimal place to me.  It was hard to get it too.

 

 

 

 

 

 

 

 

 

 

 

 

(c) John H. Mathews 2003