[Ipopt] Failed in restoration phase

[Stefan Vigerske]

Hi,

try to run the derivative checker to test if your implementation of 
grad_f is correct:
http://www.coin-or.org/Ipopt/documentation/node30.html

Stefan

On 07/23/2013 11:42 AM, 八重雾中渡 wrote:
> Hi,
> I try an example to test if scaling works and got a LOCALLY_INFEASIBLE error in my program.
> Then I change the ipopt example to test it.
> When I compute the grad_f with the partial derivative equation, I get a RESTORATION_FAILED error.
> However, the problem solved when I use the finite-difference method, whether the forward difference or central difference.
> The difference of them in central difference is less then 1e-9.
>
>
> This is the problem:
> max: f(x0, x1, x2, x3) = 0.7 * (x1 + x2 + x3 - 175000) * (x0 * x0 + 3 * x0 + 3)
> 0.5      <= x0 <= 1.5
> 45000 <= x1 <= 250000
> 10000 <= x2 <= 125000
> 5000   <= x3 <= 75000
>
>
> The following is my option:
> tol : 1e-6
> mu_strategy : "adaptive"
> hessian_approximation : "limited-memory"
> obj_scaling_factor : -1
>
>
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>