Evolutionary computing tools and machine learning techniques are efficient in antenna design. Broadly, there are two types in antenna design. Single element antenna modelling and the other is array synthesis. This is my perception and please do not compare with others.
In array design you can directly apply these non-deterministic tools directly for obtaining desired patterns. You can go through the following ... which is just an example.
Whereas coming to the antenna modeling, the interest is gain, S11 and other radiation features. Generally, the simulation is carried out using EM tools like HFSS, CST, FEKO etc...
It is even possible to employ these computing tools and techniques while modeling an antenna. In some tools like CST and FEKO these techniques are available in the integrated platform. Or otherewise you can take their API and connect to their engine using MATLAB.
Further, if you need any help please go through the literature available on my page and let me know.