How to optimize a simple function with optimize()?

[Richard Braeuer]

I am trying to run a maximum-likelihood estimation in Mata using optimize(). However, I cannot get the maximum value of even quite simple functions, so there must be a programming error with my code. I have produced a minimum working example below, where I replaced my original function with a simple x^2, just to understand what is going on:

mata
mata clear
end




mata

function logit_mata2(costs_guess)
{
printf("control")
ll = -(costs_guess)^2

return(ll)


}

logit_mata2(10)

S2 = optimize_init()
optimize_init_evaluator(S2, &logit_mata2())
optimize_init_params(S2, 10 )
result = optimize(S2)

end

I get the following error, which I have difficulty making any sense of:

logit_mata2(): 3001 expected 1 arguments but received 5
opt__calluser0_d(): - function returned error
opt__d0_calluser(): - function returned error
deriv__call1user_d(): - function returned error
_deriv__compute_value(): - function returned error
_deriv(): - function returned error
opt__eval_nr_d0(): - function returned error
opt__eval(): - function returned error
opt__looputil_iter0_common(): - function returned error
opt__looputil_iter0_nr(): - function returned error
opt__loop_nr(): - function returned error
opt__loop(): - function returned error
optimize(): - function returned error
<istmt>: - function returned error

In my mind, I have not declared any arguments exept the one parameter for which I would like the optimal value, so I do not understand what this represents

Thank you in advance!

Best,

Richard

[Paul Rathouz]

HI -- So I think the function you want to optimize has to have a specific set of input parameters and has to be a -void- function. Here is some code just recently used to determine the shape and scale of the inverse gamma distribution, where I want to know the values of those parameters that yield prespecified [by me] quantiles.

----

// Find needed shape and scale
void fnToMax(todo, p, v, g, H)
{
// Solve to get prUp=0.95 and prLo=0.05
// 5th percentile should be 1/19
// 95th percentile should be 3
// p[1]=shape and p[2]=scale
prLo=1-gammap(p[1],p[2]/(1/19))
prUp=1-gammap(p[1],p[2]/3)
v=-((prUp-.95)^2+(prLo-0.05)^2)
}
S = optimize_init()
optimize_init_evaluator(S, &fnToMax())
optimize_init_params(S,(.1,0.5))
gamParms = optimize(S)
gamParms

----