The Computable Differential Equation Lecture ... - Bruce E. Shapiro
The Computable Differential Equation Lecture ... - Bruce E. Shapiro The Computable Differential Equation Lecture ... - Bruce E. Shapiro
54 CHAPTER 3. APPROXIMATE SOLUTIONS as n → ∞. Hence the sequence y n (t) converges uniformly to a continuous function y(t). Consequently, y n (t) → y(t) uniformly, and since f(t, y) is continuous on a rectangle R, the sequence f(t, y n (t)) → f(t, y(t)) uniformly on R. Hence the sequence of integrals uniformly on R. ∫ t t 0 f(x, y n (x))dx → ∫ t t 0 f(x, y(x))dx (3.96) We now integrate both sides of inequality 3.93 from t 0 to t. ∫ t [ ] ∫ dyn (x) t ∣[ ∣∣∣ ∣ − f(x, y n (x)) dx t 0 dx ∣ ≤ dyn (x) ∣∣∣ − f(x, y n (x))]∣ dx (3.97) t 0 dx ∫ t ≤ ɛ n dx (3.98) t 0 ≤ ɛ n |t 0 − t| (3.99) ≤ ɛ n h (3.100) By the fundamental theorem of calculus, since the y n are continuous, we can also integrate the first term on the left side of the inequality, ∫ t ∣ y n(t) − y n (t 0 ) − f(x, y n (x))dx ∣ ≤ ɛ nh (3.101) t 0 Taking the limit of both sides of the equation as n → ∞, ∫ t y(t) = y(t 0 ) + f(x, y(x))dx (3.102) t 0 We observe that this function satisfies the initial value problem. Hence a solution exists. When we defined the ɛ-approximate solution we only required that it satisfy the initial value problem, namely, the differential equation and the initial condition. We did not set any restrictions on whether or not the ”approximation” matched the real solution in any way whatsoever. The following theorem tells us that we can construct an ɛ-approximation that is arbitrarily close to the actual solution. In other words, it is possible to use this technique to find a solution to the IVP that matches the true solution with any arbitray accuracy. Theorem 3.7. Cauchy-Euler Approximation Theorem. Suppose that y(t) is an exact solution of the initial value problem y ′ = f(t, y) (3.103) y ′ (t 0 ) = y 0 (3.104) Math 582B, Spring 2007 California State University Northridge c○2007, B.E.Shapiro Last revised: May 23, 2007
CHAPTER 3. APPROXIMATE SOLUTIONS 55 where f(t, y) ∈ L(y; K)(D). Let ỹ(t) be an ɛ-approximate solution to y(t) satisfying ỹ ′ (t 0 ) = y 0 for |t − t 0 | ≤ h. Then there exists a constant N > 0, independent of ɛ, such that |ỹ(t) − y(t)| ≤ ɛN (3.105) for |t − t 0 | ≤ h. Proof. From the fundamental inequality, |ỹ(t) − y(t)| ≤ e K|t−t0| |ỹ(t 0 ) − y(t 0 )| + ɛ + 0 K ≤ e Kh |y 0 − y 0 | + ɛ ( ) e Kh − 1 K ≤ ɛ ( ) e Kh − 1 K ( ) e K|t−t0| − 1 (3.106) (3.107) (3.108) So if we let N = 1 K ( ) e Kh − 1 (3.109) the conclusion to the theorem follows immediately. Theorem 3.8. Let D be an arbitrary bounded domain, f ∈ L(y; K)(D). Let y(t) be an exact solution of the IVP y ′ = f(t, y), y(t 0 ) = y 0 that is also defined for t = t 1 . The the value of y(t 1 ) may be determined to any desired degree of accuracy, using Euler’s method, but using a sufficiently small step size. Proof. Let d be the minimum distance between the curve y(t) and the boundary of D. If η is any number 0 < η < d then the strip S = {(t, y)|t 0 ≤ t ≤ t 1 , |y − y(t)| ≤ η} (3.110) lies entirely in D.Let E be the desired error. Choose ɛ so that ( ) Kη ɛ = min E, e K|t 1−t 0 | − 1 By equation 3.108 |ỹ(˜t) − y(˜t)| ≤ ɛ K ( ) e K|˜t−t 0 | − 1 (3.111) (3.112) ≤ η eK|t 1−t 0 | − 1 e K|˜t−t 0 | − 1 (3.113) < η (3.114) Thus ỹ(t) is within the strip S. Within S, the conditions of the Fundamental Inequality are met, so we can continue to approximate y by ỹ.Hence ỹ(t) is defined ⇐= up to at least t = t 1 with the desired error. c○2007, B.E.Shapiro Last revised: May 23, 2007 Math 582B, Spring 2007 California State University Northridge
- Page 9 and 10: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 11 and 12: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 13 and 14: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 15 and 16: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 17 and 18: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 19 and 20: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 21 and 22: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 23 and 24: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 25 and 26: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 27 and 28: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 29 and 30: CHAPTER 1. CLASSIFYING THE PROBLEM
- Page 31 and 32: Chapter 2 Successive Approximations
- Page 33 and 34: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 35 and 36: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 37 and 38: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 39 and 40: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 41 and 42: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 43 and 44: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 45 and 46: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 47 and 48: CHAPTER 2. SUCCESSIVE APPROXIMATION
- Page 49 and 50: Chapter 3 Approximate Solutions 3.1
- Page 51 and 52: CHAPTER 3. APPROXIMATE SOLUTIONS 45
- Page 53 and 54: CHAPTER 3. APPROXIMATE SOLUTIONS 47
- Page 55 and 56: CHAPTER 3. APPROXIMATE SOLUTIONS 49
- Page 57 and 58: CHAPTER 3. APPROXIMATE SOLUTIONS 51
- Page 59: CHAPTER 3. APPROXIMATE SOLUTIONS 53
- Page 63 and 64: CHAPTER 3. APPROXIMATE SOLUTIONS 57
- Page 65 and 66: CHAPTER 3. APPROXIMATE SOLUTIONS 59
- Page 67 and 68: CHAPTER 3. APPROXIMATE SOLUTIONS 61
- Page 69 and 70: CHAPTER 3. APPROXIMATE SOLUTIONS 63
- Page 71 and 72: CHAPTER 3. APPROXIMATE SOLUTIONS 65
- Page 73 and 74: CHAPTER 3. APPROXIMATE SOLUTIONS 67
- Page 75 and 76: Chapter 4 Improving on Euler’s Me
- Page 77 and 78: CHAPTER 4. IMPROVING ON EULER’S M
- Page 79 and 80: CHAPTER 4. IMPROVING ON EULER’S M
- Page 81 and 82: CHAPTER 4. IMPROVING ON EULER’S M
- Page 83 and 84: CHAPTER 4. IMPROVING ON EULER’S M
- Page 85 and 86: CHAPTER 4. IMPROVING ON EULER’S M
- Page 87 and 88: CHAPTER 4. IMPROVING ON EULER’S M
- Page 89 and 90: CHAPTER 4. IMPROVING ON EULER’S M
- Page 91 and 92: CHAPTER 4. IMPROVING ON EULER’S M
- Page 93 and 94: CHAPTER 4. IMPROVING ON EULER’S M
- Page 95 and 96: Chapter 5 Runge-Kutta Methods 5.1 T
- Page 97 and 98: CHAPTER 5. RUNGE-KUTTA METHODS 91 w
- Page 99 and 100: CHAPTER 5. RUNGE-KUTTA METHODS 93 w
- Page 101 and 102: CHAPTER 5. RUNGE-KUTTA METHODS 95 5
- Page 103 and 104: CHAPTER 5. RUNGE-KUTTA METHODS 97 F
- Page 105 and 106: CHAPTER 5. RUNGE-KUTTA METHODS 99 T
- Page 107 and 108: CHAPTER 5. RUNGE-KUTTA METHODS 101
- Page 109 and 110: CHAPTER 5. RUNGE-KUTTA METHODS 103
CHAPTER 3. APPROXIMATE SOLUTIONS 55<br />
where f(t, y) ∈ L(y; K)(D). Let ỹ(t) be an ɛ-approximate solution to y(t) satisfying<br />
ỹ ′ (t 0 ) = y 0 for |t − t 0 | ≤ h. <strong>The</strong>n there exists a constant N > 0, independent of ɛ,<br />
such that<br />
|ỹ(t) − y(t)| ≤ ɛN (3.105)<br />
for |t − t 0 | ≤ h.<br />
Proof. From the fundamental inequality,<br />
|ỹ(t) − y(t)| ≤ e K|t−t0| |ỹ(t 0 ) − y(t 0 )| + ɛ + 0<br />
K<br />
≤ e Kh |y 0 − y 0 | + ɛ ( )<br />
e Kh − 1<br />
K<br />
≤ ɛ ( )<br />
e Kh − 1<br />
K<br />
( )<br />
e K|t−t0| − 1<br />
(3.106)<br />
(3.107)<br />
(3.108)<br />
So if we let<br />
N = 1 K<br />
( )<br />
e Kh − 1<br />
(3.109)<br />
the conclusion to the theorem follows immediately.<br />
<strong>The</strong>orem 3.8. Let D be an arbitrary bounded domain, f ∈ L(y; K)(D). Let y(t)<br />
be an exact solution of the IVP<br />
y ′ = f(t, y), y(t 0 ) = y 0<br />
that is also defined for t = t 1 . <strong>The</strong> the value of y(t 1 ) may be determined to any<br />
desired degree of accuracy, using Euler’s method, but using a sufficiently small step<br />
size.<br />
Proof. Let d be the minimum distance between the curve y(t) and the boundary of<br />
D. If η is any number 0 < η < d then the strip<br />
S = {(t, y)|t 0 ≤ t ≤ t 1 , |y − y(t)| ≤ η} (3.110)<br />
lies entirely in D.Let E be the desired error. Choose ɛ so that<br />
(<br />
)<br />
Kη<br />
ɛ = min E,<br />
e K|t 1−t 0 | − 1<br />
By equation 3.108<br />
|ỹ(˜t) − y(˜t)| ≤ ɛ K<br />
( )<br />
e K|˜t−t 0 | − 1<br />
(3.111)<br />
(3.112)<br />
≤ η eK|t 1−t 0 | − 1<br />
e K|˜t−t 0 | − 1<br />
(3.113)<br />
< η (3.114)<br />
Thus ỹ(t) is within the strip S. Within S, the conditions of the Fundamental<br />
Inequality are met, so we can continue to approximate y by ỹ.Hence ỹ(t) is defined ⇐=<br />
up to at least t = t 1 with the desired error.<br />
c○2007, B.E.<strong>Shapiro</strong><br />
Last revised: May 23, 2007<br />
Math 582B, Spring 2007<br />
California State University Northridge