Today’s Progress

Other than my previously mentioned analysis of the transition function, I am happy to announce the following substantive progress:

I consolidated the functions “transition”, “uniformTransition”, and “gaussianTransition” into a single “transition” function.

In so doing, the following improvements were accrued:

1. Smaller code base for the transition functionality.
2. Option to supply a sample from a circle of given radius to the transition functions rather than making a new one every time a transition needs to be computed.
3. No more problems with stray optional arguments (admittedly, there is no way to supply them).

What needs to be done for this function are the usuals for my work: argument checking and documentation. I need to rewrite the documentation for the transition function and write many, many lines of argument checking code.

Next Steps:

1. Modify the IMA code to reflect the new transition function
2. Test the IMA code with the new transition function using both uniform and Gaussian proposals/transitions.

The big question is whether or not it will all work out of the box. I predict the uniform functionality will work and the Gaussian will still be, for lack of a better word, confusing.