Numerical Methods for Eng [ENGR 391] [Lyes KADEM 2007] IX. Multiple roots A multiple root corresponds to a point where a function is tangent to the x-axis: Double root Triple root f(x)=(x-2) (x-2) (x-3) f(x)= (x-2) (x-2) (x2) (x-3) In general, odd multiple roots cross the x-axis whereas even ones do not. The presence of a multiple root is very problematic for the methods defined previously because: - - There is no change in the sign of the function, so the bracketing methods cannot be applied. f x but also f ' x goes to zero, and Newton-Raphson and secant method contain f ' x in the denominator. However, it can be demonstrated theoretically that f x → 0 ' before f x → 0. Therefore, a zero check for f x must be incorporated into the ' computer program and the computation can be terminated before f x reaches zero. For multiple roots, Newton-Raphson and the secant methods are linearly rather than quadratically convergent. For open methods, a new approach has to be used to overcome the limitations introduced by multiple roots. The new formula can be written under the following form: xi1 xi m f xi f ' xi Where (m) is the multiplicity of the root (m=2 for double; m=3 for triple, …). The problem here is that you have to know the multiplicity of the root. Another approach is to define a new function U(x) as the ratio of the function to its derivative: U x f x f ' x This function has the same root as f x , it is then substituted into the original formula of NewtonRaphson: xi1 xi U xi U ' xi with U ' x f ' x f ' x f x f '' x f x ' Therefore, Roots of Equations 34 2 Numerical Methods for Eng [ENGR 391] [Lyes KADEM 2007] xi1 xi f xi f ' xi f x f x f x 2 ' '' i i i However, the price to pay for this approach is more computational effort. For the secant method the formula will be: xi1 xi U xi xi1 xi U xi1 U xi X. Systems of non-linear equations Usually to solve a system of non-linear equations, we will use an extension of open methods. An example of a system of non-linear equations is: f1 x1 , x2 x12 x22 8 x1 4 x2 11 0 f 2 x1 , x2 x12 x22 20 x1 75 0 X.1. Fixed point iteration for systems of non-linear equations One of the most important drawbacks of the fixed iteration method is that the convergence of the method is dependent on how the equations are formulated. It can be shown that sufficient convergence criteria for two equations are: f1 f 1 1 x1 x2 and f 2 f 2 1 x1 x2 This represents a very restrictive criteria and that’s why fixed point iteration method is not used to solve systems of non-linear equations. X.2. Newton-Raphson for systems of non-linear equations The Newton-Raphson formula is the following: xi1 xi f xi f ' xi This formula can be obtained using Taylor series expansion [assignment II]. We can do the same approach for a system of equations, but considering a Taylor series that account for the presence of both variables: Roots of Equations 35 Numerical Methods for Eng [ENGR 391] f1i1 f1i x1i1 x1i f1i [Lyes KADEM 2007] x2i1 x2i x1 f1i x2 ... and f 2i1 f 2i x1i1 x1i f 2i x1 x2i1 x2i f 2i x2 ... For the root estimate f1i 1 and f 2 i 1 must be equal zero. Therefore: f1i f f f x1i1 1i x2i1 f1i x1i 1i x2i 1i x1 x2 x1 x2 and f 2i f f f x1i1 2i x2i1 f 2i x1i 2i x2i 2i x1 x2 x1 x2 Finally; x1i1 x1i f1i f 2i x2 f1i f 2i x1 x2 x2i1 x2i f 2 i f1i f 2 i f 2i x1 f 2i x1 x2 x2 f1i f 2i x2 x1 f1i f1i f1i x1 f1i f 2i x2 x1 The denominator is called the Jacobian. This is the two equation version for Newton-Raphson method. This method is also highly dependent upon your initial guess. For the one equation problem, plotting the function was a good reflex to choose the initial guess. However, this is not possible with the system of equations. Therefore, we will mostly use a trial an error method or consider the physical system being modeled. Note also that the Newton-Raphson method can be generalized to solve N simultaneous equations. To solve this system, we have to write it under matrix form: J x f Where x1,2,3,… = x1,2,3,…(i+1) - x1,2,3,…(i) Roots of Equations 36 Numerical Methods for Eng [ENGR 391] [Lyes KADEM 2007] Example: Solve the following system using Newton-Raphson method: f1 x1 , x2 x12 x22 8 x1 4 x2 11 0 f 2 x1 , x2 x12 x22 20 x1 75 0 By tacking a starting point as (x1=2; x2=4) and 10 5 . Roots of Equations 37