simple numerical solve (2-variables!!)

solve
nsolve

asked 2012-09-17 07:01:13 +0200

mattias
21 ●2 ●3 ●6

updated 2012-09-20 08:54:48 +0200

Hello,

Thanks for the help, I'm trying to move from Mathematica to Sage, but I'm still having some trouble with the basics. Specifically, I'm moving from 1-var to 2-var numerical optimization/solver. I want to optimize function g or, equivalently, solve the system of first-order conditions, f. So far, I have not been able to get any of the scipy routines to work. Is there a simple way to do this? (Note: the objective function is concave wrt to both arguments.)

Thanks! jv

Constants

y_a = 50 y_b = 50 x = 40 alpha_a = .2 alpha_b = .2

Functions

v(n)= n^.5 v1(n) = derivative(v(n),n) u(n)= n^.5 u1(n) = derivative(u(n),n)

Variables

var('g_a') var('x_a')

Optimization (over [0, x])

g = (v(g_a+alpha_b(x-g_a))-u(y_a)+u(y_a-x_a))(v(x-g_a+alpha_a*(g_a))-u(y_b)+u(y_b-(x-x_a)))

f1 = ((1-alpha_b)v1(g_a+alpha_b(x - g_a)))/((1-alpha_a)v1(x-g_a+alpha_a(g_a)))==(v(g_a+alpha_b(x-g_a))-u(y_a)+u(y_a-x_a))/(v(x-g_a+alpha_a(g_a))-u(y_b)+u(y_b-(x-x_a)))

f2 = (u1(x_a)/u1(x-x_a)) == (v(g_a+alpha_b(x-g_a))-u(y_a)+u(y_a-x_a))/(v(x-g_a+alpha_a(g_a))-u(y_b)+u(y_b-(x-x_a)))

f(g_a, x_a) = (f1, f2)

edit retag flag offensive close merge delete

Comments

Did you completely change your question? If you have another question to ask, please revert this one to its first version and ask your second question as a separate topic.

niles ( 2012-09-20 09:43:00 +0200 )edit

Hi mattias. If you restore the first question and ask the second as a separate one I can show how to obtain (20,20) as the answer

achrzesz ( 2012-09-20 14:56:07 +0200 )edit

add a comment

2 Answers

Sort by » oldest newest most voted

answered 2012-09-17 10:01:35 +0200

niles
4519 ●23 ●72 ●132 http://nilesjohnson.net/

updated 2012-09-17 10:05:48 +0200

What function are you using to do this? simplify? Showing us a minimal example might be helpful.

If you're not using it already, I recommend find_root for numerical solutions to 1-variable equations.

As far as I can tell, Maxima is used primarily for symbolic manipulations in Sage, and other things are used for numerical optimization. So if you're using a function that uses Maxima, you might only get symbolic output.

edit flag offensive delete link

Comments

Thanks, this is exactly what I was looking for. I was using solve with a rather messy function. After a bit of troubleshooting I was able to get find_root to work for me.

mattias ( 2012-09-17 12:57:16 +0200 )edit

add a comment

answered 2012-09-17 07:52:12 +0200

achrzesz
2728 ●5 ●25 ●59

updated 2012-09-17 12:46:54 +0200

var('g_a')
eq=g_a == -31375155/31496372*sqrt(-4/5*g_a + 1) + 31375155/31496372*sqrt(4/5*g_a + 8) - 35/8
solve(eq,g_a,to_poly_solve=True)
[g_a == (-35/8)]
bool(eq.subs(g_a=-35/8))
#True

edit flag offensive delete link

add a comment

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

simple numerical solve (2-variables!!)

Constants

Functions

Variables

Optimization (over [0, x])

Comments

2 Answers

Comments

Your Answer

Question Tools

Stats

Related questions

simple numerical solve (2-variables!!) edit

Constants

Functions

Variables

Optimization (over [0, x])

Comments

2 Answers

Comments

Your Answer

Question Tools

Stats

Related questions

simple numerical solve (2-variables!!)