Re: RSolve and complex solutions

*To*: mathgroup at smc.vnet.net*Subject*: [mg54711] Re: [mg54693] RSolve and complex solutions*From*: Andrzej Kozlowski <akoz at mimuw.edu.pl>*Date*: Mon, 28 Feb 2005 03:26:57 -0500 (EST)*References*: <200502270629.BAA25405@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

On 27 Feb 2005, at 07:29, Skirmantas wrote: > It may be a basic math question, but it would be helpful if somebody > could briefly comment on it. If I have a system of recursion equations > with real coefficients and solve it numerically step-by-step, all my > solutions at all time steps are (obviously) real. If, however, I use > RSolve to get the general formulas for these solutions, some of them > become complex numbers. My understanding is the imaginary parts of > these numbers are due to rounding errors and the actual solutions are > only the real parts. Am I right? > > > Are all your equations linear? Otherwise the answer is "obviously no". A trivial example that satisifes all your stated conditions is RSolve[a[n]^2 + 1 == 0, a[n], n] {{a[n] -> -I, a[n] -> I}} Andrzej Kozlowski

**References**:**RSolve and complex solutions***From:*skirmantas.janusonis@yale.edu (Skirmantas)