* Improve parameters when dimension of NelderMead is one. In particular the shrink parameter is set to 0, which causes the algorithm to fail to converge.
Fixes#542.
* Add test for 1-D NelderMead
* optimize: Remove Local function
This change removes the Local function. In order to do so, this changes the previous LocalGlobal wrapper to LocalController to allow Local methods to be used as a Global optimizer. This adds methods to all of the Local methods in order to implement GlobalMethod, and changes the tests accordingly. The next commit will fix all of the names
* optimize: Change Settings to allow InitialLocation
This modifies Settings to allow specifying an initial location and properties of the function (value, gradient, etc.). This allows to work with local optimizers that are seeded with initial settings. This has two fields that must be specified, InitX and InitValues. Ideally this would only be one location, but the difficulty is that the default value of the function is 0. We either must require the user to specify it is set (in this case that InitValues is non-zero), or require the user to change the default value away if it is not set. The former seems much safer.
* optimize: remove Local implementation and replace with a call to Global
This PR starts the process described in #482. It removes the existing Local implementation, replacing with a function that wraps Method to act as a GlobalMethod. This PR also adds a hack to fix an inconsistency with FunctionConverge between Global and Local (and a TODO to make it not a hack in the future)
