Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c NUMERICAL DIFFERENTIATION AND INTEGRATION 1 NUMERICAL DIFFERENTIATION Numerical differentiation is used to determine one of the derivatives of a given function of the form y= f(x) at the point x= xk. The derivative of a function at x is defined as ( f'(x) = ) ( ) To be able to find a derivative numerically, one could make f’(x) ( ) finite to give, ( ) Knowing the value of x at which you want to find the derivative of f x, we choose a value of x to find the value of f x. To estimate the value of f x, three such approximations are Suggested : Forward Divided Difference, Backward Divided Difference and Central Difference Approximation. Continuous Function a. Forward Difference Approximation of the First Derivative From differential calculus, we know ( f’(x) ) ( ) For a finite, f’(x) ( ) ( ) The above is the forward divided difference approximation of the first derivative. It is called forward because you are taking a point ahead of x. To find the value of f’(x) at x= x i, we may choose another point ahead at x=xi+1. This gives f’( ) ( ( f’(x)= ( )) ( ) ) ( ) 1 , since Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c Graphical representation of forward difference approximation of first derivative Backward Difference Approximation of the First Derivative We know ( f’(x) ) ( ) For a finite, f’(x) If is ( ) ( ) chosen as a negative number, ( ) ( ) ( ) ) ( ) ( ) f’(x) = This is a backward difference approximation as you are taking a point backward from x. To find the value of f xat x xi , we may choose another point x behind as x=xi-1 . This gives f’(x)= ( ) ( 2 ( ) , Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c Graphical representation of backward difference approximation of first derivative . Forward Difference Approximation from Taylor Series Taylor’s theorem says that if you know the value of a function f (x) at a point xi and all its derivatives at that point, provided the derivatives are continuous between xi and xi+1 , then f ( xi1 ) f ( xi ) f ( xi )( xi1 xi ) f ( xi1 ) f ( xi ) f ( xi )(x) 1 f ( xi )( xi1 xi ) 2 ... 2! 1 f ( xi )(x) 2 ... 2! f ( xi1 ) f ( xi ) 1 f ( xi )(x) ... x 2! f ( xi1 ) f ( xi ) f ( xi ) O(x) x f ( xi ) The term O(x) shows that the error in the approximation is of the order of x . As you can see, both forward and backward divided difference approximations of the first derivative are accurate on the order of O(x). To get a better approximation of the first derivative, we employ another method called central difference approximation of the first derivative. From the Taylor series f ( xi1 ) f ( xi ) f ( xi )(x) 1 1 f ( xi )(x) 2 f ''' ( xi )(x) 3 ...(1) 2! 3! and 1 1 f ( xi )(x) 2 f ''' ( xi )(x) 3 ...( 2) 2! 3! Subtracting Equation (2) from Equation (1) 2 f ( xi1 ) f ( x x1 ) f ( xi )(2x) f ''' ( xi )(x) 3 ... 3! f ( xi1 ) f ( xi 1 ) 1 ''' f ( xi ) f ( xi )(x) 2 ... 2x 3! f ( xi1 ) f ( xi 1 ) f ( xi ) O(x) 2 2x f ( xi1 ) f ( xi ) f ( xi )(x) Hence, showing that we have obtained a more accurate formula as the error is of the order of O(x)2. 3 Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c Graphical representation of central difference approximation of first derivative. Example 1: The velocity of a rocket is given by: V (t)= 2000 * +-9.8t, 0 ≤ t ≤ 30. Where: V is given in m/s and t is given in seconds. At t=16s, and a step size oft=2s . b. Use the forward and backward difference approximation of the first derivative of V(t) to calculate the acceleration. c. Find the exact value of the acceleration of the rocket. d. Find the absolute relative true error for both the forward and backward approximations. SOLUTION Forward difference method a(t)= ( ) ( ) but ti=16s, t=2s , a(16)= V (18) = 2000 * ( ) ( ) ( ) ( ) +-9.8(18), =453.02 m/s V (16) = 2000 * +-9.8(16), =392.07m/s a(16)= ( ) ( ) = 4 =30.474m/s2 Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c Backward difference method a(t) ( ) ( ) , but ti=16s, t=2s , a(16)= ( ) ( ) * V (16 = 2000 ( +-9.8(16), ) =392.07 m/s * V (14) = 2000 ( +-9.8(14), ) =334.24m/s ( a(16)= ) ( ) =28.915m/s2 )= b. The exact value of a (16) can be calculated by differentiating V (t)= 2000 +-9.8t, 0 ≤ t ≤ 30. * As and = ( ) Knowing that , a(t)= ( ) a(t)= 2000( [ ]) -9.8, )( =2000( )* ( ) +( = ( a(16)= ( ) ) =29.674m/s2 c. The absolute relative true error is: Forward: | | | | | | | | =2.6967% Backward: 5 ) - 9.8, Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c | | | | Example 2: Repeat example 1 for the central difference approximation of the first derivative. V (t i 1 ) V (t i 1) 2t but ti=16s, t=2s , a (t i ) a (16) V (18) = 2000 * ( ) ( ) V (18) V (14) 2( 2) +-9.8(18), =453.02 m/s V (14) = 2000 * +-9.8(14), =334.24m/s 453.02 334.24 = 29.694m/s2 2(2) (b) The exact value of the acceleration at from Example 1 is s t=16s a(16) a(16) =29 .674m/s2 The absolute relative true error for the answer in part (a) is | | | | From the results in example 1 (forward and backward) and example 2 (central difference), it is clearly shown that the relative approximate error for the central difference approximation is very small, showing that it gives more accurate result. This is because the order of accuracy is proportional to the square of the step size. Finite Difference Approximation of Higher Derivatives One can also use the Taylor series to approximate a higher order derivative. For example, to approximate f''(x) , the Taylor series is 6 Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c f ( xi 2 ) f ( xi ) f ( xi )(2x) 1 1 f ( xi )(2x) 2 f ''' ( xi )(2x) 3 ...(3) 2! 3! where, x i 2 xi 2x f ( xi 1 ) f ( xi ) f ( xi )(x) 1 1 f ( xi )(x) 2 f ''' ( xi )(x) 3 ...( 4) 2! 3! where, x i-1 xi x subtracting 2 times Equation (4) from Equation (3) gives f ( xi 2 ) 2 f ( xi 1 ) f ( xi ) f ( xi )(x) 2 f ''' ( xi )(x) 3 ... f ( xi 2 ) 2 f ( xi 1 ) f ( xi ) f ''' ( xi )(x) .... 2 ( x) f ( xi 2 ) 2 f ( xi 1 ) f ( xi ) f ' ' ( xi ) O( x) ( x) 2 f ' ' ( xi ) 2. NUMERICAL INTEGRATION Integration is the process of measuring the area under a function plotted on a graph. A wide variety of numerical methods have been developed to simplify an integral. We will discuss four well-known methods: the trapezoidal rule, Simpson’s 1/3rd Rule, Gauss Quadrature Rule, and Romberg Integration. Integrating discrete functions Sometimes, the function to be integrated is given at discrete data points, and the area under the curve is needed to be approximated. Here, we will discuss the integration of such discrete functions, ∫ ( ) Where: f(x)- is called the integrand and is given at discrete value of x , a- lower limit of integration b -upper limit of integration Multiple methods of integrating discrete functions are discussed below. 2.1 Rectangular Rule Approximate the integration, ∫ ( ) that is the area under the curve by a series of rectangles as shown below. The base of each of these rectangles is x=(b-a)/n and its height can be expressed as f(xi*) where xi* is the midpoint of each rectangle and n is number of rectangle. b f ( x)dx f ( x *)x f ( x *)x .. f ( x *)x a 1 2 n x[ f ( x1*) f ( x2 *) .. f ( xn *)] 7 Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c 2.2 Trapezoidal Rule The rectangular rule can be made more accurate by using trapezoids to replace the rectangles as shown. A linear approximation of the function locally sometimes works much better than using the averaged value like the rectangular rule does. b x x x f ( x)dx 2 [ f (a ) f ( x )] 2 [ f ( x ) f ( x )] .. 2 [ f ( x ) f (b)] a 1 1 1 1 x[ f (a ) f ( x1 ) .. f ( xn 1 ) f (b)] 2 2 8 2 n 1 Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c 2.3 Simpson’s Rule Still, the more accurate integration formula can be achieved by approximating the local curve by a higher order function, such as a quadratic polynomial. This leads to the Simpson’s rule and the formula is given x b f ( x)dx 3 [ f (a ) 4 f ( x ) 2 f ( x ) 4 f ( x ) .. 1 a 2 3 ..2 f ( x2 m2 ) 4 f ( x2 m1 ) f (b)] It is to be noted that the total number of subdivisions has to be an even number in order for the Simpson’s formula to work properly. Example 5 Integrate f ( x ) x 3 between x 1 and x 2. 2 1 4 2 1 4 4 3 1 x dx= 4 x |1 4 (2 1 ) 3.75 Using 4 subdivisions for the numerical integration: x= 2-1 0.25 4 Rectangular rule: 2 x dx 3 1 x[ f (1.125) f (1.375) f (1.625) f (1.875)] 0.25(14.9) 3.725 Trapezoidal Rule 2 x dx 3 1 1 1 f (1) f (1.25) f (1.5) f (1.75) f (2)] 2 2 0.25(15.19) 3.80 x[ 9 Numerical Differentiation And Integration For 3rd Year Students of Dambi Dollo University,2012E.c Simpson’s Rule 2 x x dx 3 [ f (1) 4 f (1.25) 2 f (1.5) 4 f (1.75) f (2)] 3 1 0.25 (45) 3.75 perfect estimation 3 10