数值分析chapter9.pptVIP

(9.1) Theorem 9.1 Suppose in (9.1) ⑴ is continuous on ⑵ satisfies Lipschitz in the variable y on D，that is, (9.2) where L is called Lipschitz constant. Then, problem (9.1) has a unique solution . Chapter 9 Numerical Solution of Ordinary Differential Equation §1 Euler’s Method Euler’s method is the most simple method for solving ODE, its precision isn’t very good. But this method provides the idea of solving ODE and based on this method, many better methods can be developed. 1.1 Euler’s Method Divide the interval into n equal parts, the step are mesh points. Expand y(xi+1) at point xi Leave out the error , we have The truncating error （9.3） Suppose yi is an approximation of y(xi)，then Taylor Expansion (9.4) (9.4) is called Euler’s method for (9.1). Numerical Differentiation When h is sufficient small, the first divided difference is an approximation of a derivative Numerical Integration Integrating equation on the interval Using rectangle formula to integration term, Thus, ,we have 1.2 Modified Euler Method Midpoint Method Using mid-rectangle formula to integration term, we have （9.5） The truncating error （9.5） is called midpoint formula. Trapezoid Method For When we use trapezoid formula to get integration, We have We have formula （9.6） The truncating error （9.6） is called trapezoid formula. In the formula, yi+1 is unknown and we have solve it. Eg 9.1 Using Euler’s method and predictor-corrector method approximate the solution of differential equation, and compare the results（h=0.1） Predictor-Corrector Method (9.7) Solution. （1）analytical solution （2）Euler’s method （3）Predictor-corrector method Euler P-C Exact Euler P-C Exact 0.1 1.100000 1.095909 1.095445 0.6 1.508966 1.485956 1.483240 0.2 1.191818 1.184096 1.183216 0.7 1.580338 1.552515 1.549193 0.3 1.277438 1.266201 1.264911 0.8 1.649783 1.616476 1.612452 0.4 1.358213 1.343360 1.341641 0.9 1.717779 1.678168 1.673320 0.5 1.435133 1.416402 1.414214 1