You may not get any choice about whether to use numbered subheadings or not in your thesis, but if you do have a choice, here are some things to think about.
I don’t like numbered headings all that much myself – well at least when it comes to numbering subsections and subsections of subsections of subsections. Telling the difference between 22.214.171.124.5 and 126.96.36.199.0 is too much work for me as a reader – I should be able to tell where we are up to without trying so hard. A document that is well formatted with clear heading levels and use of white space, and also uses appropriate transitions to communicate metastructure doesn’t need all these numbers to tell me where I am up to.
In this post Pat Thomson makes a similar point, saying that the numbering subsections and subsections of subsections breaks up the text into a series of points or lists and this can affect the flow of writing. She has also noticed a tendency for less informative headings if subsection numbering is used. In both cases this points to writers using the numbers to tell readers something about the writing structure when it would be better for the words to do this.
Why is it better for the words to do this? Because with words you can show readers your deep connected understanding of the subject you are discussing, and how the parts relate to each other. You can show them how each thing you say connects to what you have been saying and are going to say next. Your thesis is making new knowledge. A list–even of new stuff–isn’t knowledge. Knowledge connects to what is already known and to knowledge yet to come. The structure of your writing can illustrates the structure of your knowledge and show readers (and examiners) that you are indeed an expert in the field of your research.