maxLik is a statistical package for maximum likelihood estimation. For example, the max of can be found like the below using newton raphson method.

> x <- seq(0, 7, .01)
> f <- function(x) { -1 * (x-2) * (x-5) }
> plot(x, f(x), type="l")
> abline(a=0, b=0)
> m <- maxNR(f, start=-5, print.level=2)
----- Initial parameters: -----
fcn value: -70 
     parameter initial gradient free
[1,]        -5               17    1
Condition number of the (active) hessian: 1 
-----Iteration 1 -----
-----Iteration 2 -----
-----Iteration 3 -----
--------------
gradient close to zero 
3  iterations
estimate: 3.5 
Function value: 2.25 
> summary(m)
--------------------------------------------
Newton-Raphson maximisation 
Number of iterations: 3 
Return code: 1 
gradient close to zero 
Function value: 2.25 
Estimates:
     estimate     gradient
[1,]      3.5 8.881784e-10
--------------------------------------------

Estimate was 3.5, and this is correct because .

Optimization and Mathematical Programming has list of packages for optimization.