First, do you have a good general purpose antenna reference text? If not, I suggest you invest in one (or several). Classic authors include Balanis (Antenna Theory, Analysis and Design), Stutzman & Thiele (Antenna Theory and Design), Johnson & Jasik (Antenna Engineering Handbook), Kraus (Antennas), and on and on... My personal choice was Constantin Balanis's text, because he devotes an entire chapter to microstrip antennas.
Second, you have to collect your tools. Do you just want to 'cut and try'? Do you want to go for 'first pass?' Or something in between. You can do a pretty good job of guesswork with a few simple equations and a spreadsheet. A mathematical tool, MathCAD, Matlab, Mathematica, or such, is a little better, because you can evaluate the necessary integrals for radiation resistance and mutual impedance (hint, hint). A linear simulator is nice to have. Field solvers are also nice, but to do any serious work takes serious money. How are you going to measure your antenna to see if it makes you happy?
Third, you need to know what you are after. Frequency? gain? polarization? bandwidth? match? pattern? (noise temperture?)
I've had pretty good success modeling patch antennas as microstrip tranmission lines, terminated with estimated fringing capacitance and 'radiation resistors.'
The first bit of info needed is "what is the bandwidth", or "what frequency range" do you need to use.
Thin antennas are small bandwidth (small frequency range) and thick antennas are wide bandwidth. Patches go from 0.1% up to 50+% bandwidth.
Switches - Pasternack or Minicircuits to start. Again, frequency and transmit power is needed.
