,calc ,a1 ,lab #9 ,9troduc;n4 ,we're pursu+ two approa*es 6-put+ >1s "r n[1 "o 9volv+ antid]ivatives & "o 9volv+ sums ( >1s ( n>r[ slices4 ,! f/ ( ^! gives rise 6! simple/ =mulas1 b x is hamp]$ 2c we c't alw take ! antid]ivative (a func;n4 ,z a result1 :5 -put+ >1s 9 r1l life1 "o is (t5 reduc$ 6a4+ up ! >1s ( slices4 ,un=tunately1 we norm,y c't get exact =mulas =! sum (! >1s (a bun* ( slices1 s9ce we don't h any grt te*nology = -put+ sums4 ,if we >e reduc$ 6a4+ up ! >1s ( slices1 !n1 we'll 2 j "w+ num]ic,y1 n analytic,y4 ,! purpose ( ? lab is 6look = ways 6m9imize \r num]ical "w & 6maximize ! a3uracy ( \r approxim,ns if we >e reduc$ 6a4+ up slices4 ,me?ods = approximat+ >1s4 #j ,suppose y want 6-pute ! >ea "u ! graph ( y .k f(x) 2t x .k #0 & x .k #3_4 ,suppose1 fur!r1 t y c't "w \ ! result analytic,y1 & t y ne$ 6-pute approximate num]ical >1s (a bun* ( slices & 6add up ! results4 ,= 3crete;s1 suppose y decide 6cut ! 9t]val @(0, 3@) 9to #4 slices4 ,!n ,i c ?9k ( at l1/ #4 me?ods y mi1s4 ,me?od #1_3 ,"r-h& rectangles4 ,9 ? me?od1 : we've us$ 9 class1 y wd approximate ;f 9 ea* slice 0a 3/ant func;n ^: heiea y're -put+ is !n %[n 9 ,figure #1_4 ,=m,y1 ! >ea y're -put+ is f(?3_/4#)?3_/4#+f(?6_/4#)?3_/4# +f(?9_/4#)?3_/4#+f(?12_/4#)?3_/4# .k ".,s%k .k #1<4]f(?3k_/4#)?3_/4#_4 ,me?od #2_3 ,left-h& rectangles4 ,"h y wd use ! same idea1 except y wd approximate ;f 9 ea* slice 0a 3/ant #a func;n ^: heiea y're -put+ is !n %[n 9 ,figure #2_4 ,=m,y1 ! >ea y're -put+ is f(?0_/4#)?3_/4#+f(?3_/4#)?3_/4# +f(?6_/4#)?3_/4#+f(?9_/4#)?3_/4# .k ".,s%k .k #0<3]f(?3k_/4#)?3_/4#_4 ,me?od #3_3 ,trapezoids4 ,9 lab la/ week1 we approximat$ ! >1s 0putt+ trapezoids 9 ea* slice4 ,we cd write d[n a new =mula = ? >ea1 b "! is a simpl] way 6?9k ab x4 ,9 ea* slice1 ! >ea (! trapezoid is midway 2t ! >ea (! left-h& rectangle &! >ea (! "r-h& rectangle4 ,! same ?+ has 6"w if we add up all ! trapezoids3 ! >ea has 6be ! av]age (! >1s (! "r-h& & left-h& rectangles4 ,me?od #4_3 ,midpo9ts4 ,a f9al me?od wd 2 6use rectangles ag1 b ? "t 6let ! heiea1 %[n 9 ,figure #3, wd 2 f(?0_/4#+?3_/8#)?3_/4#+f(?3_/4# +?3_/8#)?3_/4#+f(?6_/4# +?3_/8#)?3_/4#+f(?9_/4# +?3_/8#)?3_/4# .k ".,s%k .k #0<3]f(?3k_/4# +?3_/8#)?3_/4#_4 ,g5]al =mulas4 ,9 g5]al1 if we want 6approximate !;a^b"f(x)dx us+ ;n slices1 !n ! "r-h& rectangle me?od wd tell u 6-pute ! sum ".,s%k .k #1e writt5 9 ,maple =m on ! atta*$ ,maple "w%eet4 #c ,example4 ,suppose y want$ 6-pute ! >ea "u ! s9e func;n 2t x .k #0 & x .k ?.p_/2#, !;0^.p_/2"sin xdx_4 ,we've n"e d"o a pro#m l ?1 s x looks l we'll h 6approximate4 (,actu,y1 ! situ,n isn't hope.s4 ,if we 2lieve t ! >ea func;n ,f0(x) satisfies ,f'0(x) .k sin x & t ,f0(0) .k #0, !n ,f0(x) .k #1-cos x_2 s ! >ea m/ 2 ,f0(?.p_/2#) .k #1-cos ?.p_/2# .k #1_4 ,? is use;l 6"k z we />t look+ at ! ]ror 9 \r approxim,ns4 ,b let's g back 6assum+ we ne$ to approximate_4) #1_4 ,2g9 0fill+ 9 ! foll[+ ta#s1 %[+ ! approxim,ns produc$ 0ea* me?od & %[+ _! ]rors f ! exact value ( #1_4 ,foll[ ! advice on ! ,maple "w%eet 6avoid ?r[+ away preci.n unnecess>ily4 #d ,approxim,n 6! >ea ,',key 6items 9 ! foll[+ ta# #a ,left rects #b ,trapezoids #c ,midpo9ts ,' -------------------------------- ;n ,"r rects #a #b #c -------------------------------- #1 -------------------------------- #10 -------------------------------- #100 -------------------------------- #1000 -------------------------------- #10,000 -------------------------------- #100,000 -------------------------------- #1,000,000 -------------------------------- #e ,]ror 9 ! approxim,n #f ,',key 6items 9 ! foll[+ ta# #a ,left rects #b ,trapezoids #c ,midpo9ts ,' -------------------------------- ;n ,"r rects #a #b #c -------------------------------- #1 -------------------------------- #10 -------------------------------- #100 -------------------------------- #1000 -------------------------------- #10,000 -------------------------------- #100,000 -------------------------------- #1,000,000 -------------------------------- #g #2_4 ,"s columns 3ta9 approxim,ns t >e too l>ge (bi7] ?an #1)1 & o!rs 3ta9 approxim,ns t >e too small (less ?an #1 )_4 ,expla9 geometric,y h[ y cd h det]m9$ 9 adv.e : columns wd 2 :4 #3_4 (a) ,look at ! ]rors 9 ! "r rectangle me?od4 ,!y %d =m a r\< patt]n4 ,>e !y propor;nal to #1_/n or to #1_/n^2, or :at8 ,if y ne$$ 6"w \ ! 9tegral to #100 digit a3uracy us+ ! "r rectangle me?od1 r\e ! ]rors (! midpo9t & trapezoid me?ods4 ,: "w$ bett]8 ,is "! any rel,n 2t ! sizes & signs (! ]rors 9 #h ! two me?ods8 #5_4 ,exploit ! 3nec;n 9 ,pro#m #4 6su7e/ a way 6-b9e ! midpo9t &! trapezoid me?od 96a new me?od m a3urate ?an ei al"o4 ,try \ ! new me?od on \r sample 9tegral1 & rep1t pro#m #3(a) )! new me?od4 ,don't quit until y h a me?od : is way bett] ?an ! midpo9t me?od4 ,j twice z gd is n 54 #6_4 ,once y h a re,y gd me?od 9 ,pro#m #5, try x \ on a quadratic func;n4 ,see if y c prove any?+4 #i