blombardi@x102c.ess.harris.com (Bob Lombardi 44139) (01/12/91)
Greetings, Now that I've decided to buy Borland's Turbo Pascal Numerical Methods Toolbox, I find that they have discontinued the product. (I have noticed this phenomenon before, and it indeed is a law of nature, derived from electro-weak theory.) :):) Does anyone have it that would like to sell it? If not, can anyone point me to a reference that explains the Laguerre algorithm that they use, so that I can write my own? The Laguerre algorithm finds all roots of polynomials from entered coefficients. I have what I thought was a good collection of post-calculus math books, but none have anything by this name. Lacking that (I get three strikes, don't I?) can anyone refer me to source code in Pascal, BASIC, or FORTRAN that finds the roots of polynomials? I can usually translate the other two languages into TP if I get code in them. Since I've posted this to several groups, I'd appreciate reply by email to save net bandwidth. Thanks, Bob Bob Lombardi WB4EHS >>>>>>> Internet: blombardi@x102c.ess.harris.com M/S 102-4826, Harris Corp GASD, P.O. Box 94000, Melbourne, FL 32902 Hobbies: ******** on hold thanks to being a gradual student in EE ****** aspiring classical pianist. Professional: electrical engineer, writer.
price@helios.unl.edu (Chad Price) (01/13/91)
blombardi@x102c.ess.harris.com (Bob Lombardi 44139) writes: >If not, can anyone point me to a reference that explains the Laguerre >algorithm that they use, so that I can write my own? The Laguerre >algorithm finds all roots of polynomials from entered coefficients. >I have what I thought was a good collection of post-calculus math books, >but none have anything by this name. See page 279 of Numerical Recipes in C by Press,Flannery,Teukolsky, and Vetterlingg (1988, Cambridge Univ Press) for explanation and source code. There is also a Fortran version of the book. chad price price@fergvax.unl.edu
oze3@quads.uchicago.edu (james daniel ozeran) (01/13/91)
In article <price.663718555@helios> price@helios.unl.edu (Chad Price) writes: > >See page 279 of Numerical Recipes in C by Press,Flannery,Teukolsky, and >Vetterlingg (1988, Cambridge Univ Press) for explanation and source >code. There is also a Fortran version of the book. > There's also a Pascal version of the book with attendant coding examples. D. Ozeran -- J. Daniel Ozeran | Depts. of Biochemistry and oze3@midway.uchicago.edu | Molecular Biology and of jozeran@biovax.uchicago.edu| Pediatrics | University of Chicago
rcollins@altos86.Altos.COM (Robert Collins) (01/15/91)
In article <5231@trantor.harris-atd.com> blombardi@x102c.ess.harris.com (Bob Lombardi 44139) writes: > >Lacking that (I get three strikes, don't I?) can anyone refer me to >source code in Pascal, BASIC, or FORTRAN that finds the roots of >polynomials? I can usually translate the other two languages into TP >if I get code in them. > I'm not familiar with the algortithm you mentioned, but I programmed a polynomial ROOT algorithm using BAIRSTOW's method in 8087 assembler. It's short, and most definitely fast -- some 28X faster than compiled QUICK-BASIC. BAIRSTOW's method uses partial derivatives to force convergence of polynomial roots. It finds all real and complex roots. Each root of a 7th degree polynomial can usually be solved in less than 10 iterations to an accuracy of 1E-6. I wrote the assember to interface to (most) any high-level language, using any memory model size (through changing compile-time switches). So I would think this algorithm could be interfaced to Turbo-Pascal...though I have never tried it. If memory serves me correctly, the algorithm has tunable parameters. For example, the size of the polynomial can be variable -- at run time, the convergence factor is also tunable, but only at compile-time. If interested, I'd be happy to send anybody a copy of the source code. -- "Worship the Lord your God, and serve him only." Mat. 4:10 Robert Collins UUCP: ...!sun!altos86!rcollins HOME: (408) 225-8002 WORK: (408) 432-6200 x4356