Let me guess what you are really tasked with
1) This is training/university task
2) The techniques you listed are those mentioned or explained
3) Your task is further study the OAT
Give me one reason why donít you simply google Orthogonal array testing? Among top 10 there is an article by Elfriede Dustin available for free:
Are you asking for the best method to chart test coverage against your automated white box or unit tests? If so then orthoganal testing is indeed a good base measure for test coverage, but it often has a hard time being scaled up for system or integration testing.
None of the methods you listed lends itself well to functional, accesibilty, or usability testing.
I do think that being flexible enough to use one or more of the different techniques depending on the situation you are in is the best way to go. We all tend to have "favored" processes or tools, but we should look objectively at what we want to do and use the combination of techniques that will work at that time.
IMHO orthogonal arrays are overly complex to figure out, and often result in over-testing unless there is an equal number of variables for each parameter. If you enjoy pain check out Orthogonal Array Selector.
If what you are testing is the interaction of dependent parameters then take a look at Pairwise Testing . There are lots of free tools (and some commercial ones) for performing combinatorial analysis of interaction between variables of dependent parameters. (I would recommend taking a look at the PICT tool because it is free, and highly customizable.)
Of course, you must realize these are tools. The value of the tool is heavily dependent on the person using the tool and their ability to understand how to use the tool properly and effectively.