thanks to everyone who helped me . On Wed, 2005-06-22 at 10:39 -0400, Stavros Macrakis wrote: > On the other hand, solving for F symbolically seems intractable in > general, since there are subexpressions of the form F*exp(F*k) and > F+exp(F), making the equations non-algebraic (and not algebraic in > exp(F), either). But perhaps there is some structure in these > particular equations that I am missing. These equations are part of a model for measuring the reliability of a software product (Musa model ) . I have to find the absolute maximum of the function G(N,F) , so i need to equal both the partial derivatives of G(N,F) to zero. The problem is, every solution of this system could be a maximum or a minimum of G . so i have to find all the solutions of the system , even numerically, , and then to see for which values the second derivative is < 0 . i' ve found another problem with maxima , under some circumstances the mnewton script crashes . see the attachment . i'm running maxima 5.9.1 on debian 3.1 , and the mnewton script version is 1.3 .
Attached file: test3.out.zip