{VERSION 4 0 "IBM INTEL LINUX22" "4.0" } {USTYLETAB {CSTYLE "Maple Input" -1 0 "Courier" 0 1 255 0 0 1 0 1 0 0 1 0 0 0 0 1 }{CSTYLE "" -1 256 "" 1 18 0 0 0 0 0 1 0 0 0 0 0 0 0 0 } {PSTYLE "Normal" -1 0 1 {CSTYLE "" -1 -1 "Times" 1 18 0 0 0 0 2 2 2 0 0 0 0 0 0 0 }0 0 0 -1 -1 -1 0 0 0 0 0 0 -1 0 }{PSTYLE "R3 Font 0" -1 256 1 {CSTYLE "" -1 -1 "10x20" 1 20 0 0 0 0 2 2 2 0 0 0 0 0 0 0 }0 0 0 -1 -1 -1 0 0 0 0 0 0 -1 0 }{PSTYLE "R3 Font 2" -1 257 1 {CSTYLE "" -1 -1 "Courier" 1 24 0 0 0 0 2 2 2 0 0 0 0 0 0 0 }0 0 0 -1 -1 -1 0 0 0 0 0 0 -1 0 }{PSTYLE "R3 Font 3" -1 258 1 {CSTYLE "" -1 -1 "Times" 1 32 0 0 0 0 2 2 2 0 0 0 0 0 0 0 }0 0 0 -1 -1 -1 0 0 0 0 0 0 -1 0 } {PSTYLE "R3 Font 4" -1 259 1 {CSTYLE "" -1 -1 "Times" 1 12 0 0 0 0 2 2 2 0 0 0 0 0 0 0 }0 0 0 -1 -1 -1 0 0 0 0 0 0 -1 0 }{PSTYLE "" 259 260 1 {CSTYLE "" -1 -1 "" 1 18 0 0 0 0 0 1 0 0 0 0 0 0 0 0 }0 0 0 -1 -1 -1 0 0 0 0 0 0 -1 0 }} {SECT 0 {EXCHG {PARA 0 "> " 0 "" {MPLTEXT 1 0 0 "" }}}{EXCHG {PARA 258 "" 0 "" {TEXT -1 61 "Integrals and Riemann Sums in One, Two, and \+ Three Variables\n" }}{PARA 258 "" 0 "" {TEXT -1 36 "Approximating sums in dimension one\n" }}{PARA 0 "" 0 "" {TEXT -1 377 "Here are some exa mples of commands that use Maple to calculate approximating\nsums for integrals in one, two, and three variables.\n\nThe first few inputs m ay look mysterious; they simply define some of the commands\nwe'll ne ed for future reference. There's no need to worry too much about\nth e precise details of how they're defined, but do watch the syntax in \+ the\nexamples.\n\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 329 "midpointsum \+ := proc (f,interval,n) local a,b,x,i,step,total;\n a := op(1,op(2 ,interval));\n b := op(2,op(2,interval));\n x := op(1,interval );\n step := (b-a)/n;\n total := 0;\n for i from 1 to n do \n total := total + evalf( subs(x=a+i*step-step/2,f) );\n od; \n total := evalf( total*step );\n end:\n" }}}{EXCHG {PARA 0 " " 0 "" {TEXT -1 484 "The previous command defined a midpoint rule Ri emann sum for functions of\none variable. Here are some examples to show how it works. The\nlast input tells how many subdivisions of th e interval of integration\nthe rule should use. We'd expect that w ith more subdivisions, the answer ought\nto get closer and closer to t he `exact' value of the integral.\n\nSuppose, e.g., that we want to in tegrate sin(x) from x=0 to x=1. The answer\nis easy to find by \+ antidifferentiation:\n\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 20 "int( si n(x),x=0..1);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 33 "\nLet's find a n umerical version:\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 9 "evalf(%);" }} }{EXCHG {PARA 0 "" 0 "" {TEXT -1 137 "\nNow let's compare this with so me approximations to the same integral,\nusing the midpoint rule. Fir st we'll use just four subdivisions:\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 32 "midpointsum( sin(x), x=0..1, 4);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 115 "\nThe answer isn't too bad an approximation, given the s mall number of subdivisions.\nBut let's bump that number up.\n" }} {PARA 0 "> " 0 "" {MPLTEXT 1 0 32 "midpointsum( sin(x), x=0..1, 8);" } }}{EXCHG {PARA 0 "> " 0 "" {MPLTEXT 1 0 34 "midpointsum( sin(x), x=0.. 1, 100);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 272 " \nThe estimates see m to behave as they should---larger numbers of subdivisions\ntend to g ive better estimates. But it's striking how well the method works \+ even\nfor small numbers of subdivisions. \n\nTry the same idea on s ome other functions and/or other intervals. " }}{PARA 0 "" 0 "" {TEXT -1 0 "" }}{PARA 259 "" 0 "" {TEXT 256 35 "Approximating sums in \+ dimension two" }}{PARA 259 "" 0 "" {TEXT -1 0 "" }}{PARA 0 "" 0 "" {TEXT -1 425 "Now let's do some approximating sums for a function f(x ,y) of two variables. The\nidea is to approximate the integral of s uch a function over a rectangle [a,b] x [c,d] .\nFirst we have to te ll Maple the two-dimensional version of the midpoint sum\nrule. (Th ink about what midpoint means in this case!) As above, there's no \nneed to worry about details of the Maple procedure below---what it d oes\nis what's important." }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 0 "" }} {PARA 0 "> " 0 "" {MPLTEXT 1 0 552 "doublemidpointsum := proc (f,inter val1,interval2,n)\n local a,b,c,d,x,y,i,j,stepx,stepy,total;\n \+ a := op(1,op(2,interval1));\n b := op(2,op(2,interval1));\n c := op(1,op(2,interval2));\n d := op(2,op(2,interval2));\n x : = op(1,interval1);\n y := op(1,interval2);\n stepx := (b-a)/n; \n stepy := (d-c)/n;\n total := 0;\n for i from 1 to n do \n for j from 1 to n do\n total := total +\n evalf( subs(x =a+i*stepx-stepx/2,y=c+j*stepy-stepy/2,f));\n od;\n od;\n \+ total := evalf( total*stepx*stepy );\n end:" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 282 "\nNotice that there are now two intervals involved- --one in the x-variable and\none in the y-variable. \n\nLet's apply the rule to the function f(x,y) = sin(x)sin(y), on the rectangle\n[ 0,1] x [0,1], using 10 subdivisions in each direction. (This produce s a grid\nof 100 squares.)\n" }}}{EXCHG {PARA 0 "> " 0 "" {MPLTEXT 1 0 54 "doublemidpointsum( sin(x)*sin(y), x=0..1, y=0..1, 10);" }}} {EXCHG {PARA 0 "" 0 "" {TEXT -1 386 "The answer is an approximation to the integral of sin(x)sin(y) over the same\nsquare. Geometrical ly, such an integral tells the volume of the solid whose base\nis on \+ the square [0,1] x [0,1] in the xy-plane, and whose ``top'' is t he surface\nz=f(x,y). Let's look at this figure, to see whether the number above seems\nlike any sort of sensible approximation to its vo lume:\n\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 51 "plot3d( sin(x)*sin(y), x=0..1, y=0..1, axes=boxed);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 204 "\n\nTake a close look. Does it seem possible that the volume is aro und 0.2?\n\nTo find the volume ``exactly,'' we can use an iterated i ntegral, i.e., we can\nintegrate in one variable and then the other:\n \n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 28 "int( sin(x)*sin(y), x=0..1); " }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 115 "The result is a function of \+ y (as it should be ---think about it).\n\nNow we can integrate in t he other variable:\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 14 "int(%,y=0.. 1);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 69 "\n\nWe could have done the same thing all in one step, if we'd wanted:\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 43 "int( int( sin(x)*sin(y), x=0..1), y=0..1 );" }}} {EXCHG {PARA 0 "" 0 "" {TEXT -1 30 "\nLet's get a decimal value:\n" }}{PARA 0 "" 0 "" {TEXT -1 0 "" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 9 "ev alf(%);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 160 "\nNotice how similar \+ that number is to the one we found using the midpoint rule. With\nfe wer subdivisions, we get less good agreement (but still not too bad): \n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 53 "doublemidpointsum( sin(x)*sin (y), x=0..1, y=0..1, 4);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 76 "Try t he same idea on some other functions f(x,y) and some other intervals. \n" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 1 "\n" }}{PARA 260 "" 0 "" {TEXT -1 37 "Approximating sums in dimension three" }}{PARA 259 "" 0 " " {TEXT -1 0 "" }}{PARA 0 "" 0 "" {TEXT -1 359 "Approximating sums wor k much the same way for a function g(x,y,z) of three variables. \nT his time, we approximate the integral of such a function over a cube\n \n [a,b] x [c,d] x [e,f] .\n\nHere is Maple's thre e-dimensional version of the midpoint sum. It's even\nmessier than t he earlier versions, and you can safely ignore its Maple code:\n" }}} {EXCHG {PARA 0 "> " 0 "" {MPLTEXT 1 0 809 "triplemidpointsum := proc ( g,interval1,interval2,interval3,n)\n local a,b,c,d,e,f,x,y,z,i,j,k ,stepx,stepy,stepz,total;\n a := op(1,op(2,interval1));\n b := op(2,op(2,interval1));\n c := op(1,op(2,interval2));\n d := o p(2,op(2,interval2));\n e := op(1,op(2,interval3));\n f := op( 2,op(2,interval3));\n x := op(1,interval1);\n y := op(1,interv al2);\n z := op(1,interval3);\n stepx := (b-a)/n;\n stepy \+ := (d-c)/n;\n stepz := (f-e)/n;\n total := 0;\n for i from 1 to n do\n for j from 1 to n do\n for k from 1 to n do\n \+ total := total + evalf( subs(x=a+i*stepx-stepx/2,\n \+ y=c+j*stepy-stepy/2,\n \+ z=e+k*stepz-stepz/2,g));\n od; od; od;\n total := evalf( tot al*stepx*stepy*stepz );\n end:\n" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 0 "" }}{PARA 0 "" 0 "" {TEXT -1 262 "Let's use the command to appr oximate the integral of the constant function g(x,y,z)=x+y+z\nover t he cube [0,1] x [0,2] x [0,3] . Note that setting n=4 here reall y amounts\nto using 4^3 = 64 subdivisions. (Setting n=10 would invo lve 1000 subdivisions!)\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 54 "triple midpointsum( x+y+z, x=0..1, y=0..2, z=0..3, 4);" }}}{EXCHG {PARA 0 " " 0 "" {TEXT -1 121 "\nThat number can't readily be interpreted geomet rically, but it does agree with what\nwe'd get by iterated integratio n:\n\n" }}{PARA 0 "> " 0 "" {MPLTEXT 1 0 47 "int( int( int( x+y+z,x=0. .1), y=0..2), z=0..3);" }}}{EXCHG {PARA 0 "" 0 "" {TEXT -1 1 "\033" }} }}{MARK "22 0 0" 1 }{VIEWOPTS 1 1 0 1 1 1803 1 1 1 1 }{PAGENUMBERS 0 1 2 33 1 1 }