Yes, it's not really linear, and with higher dilutions, you will likely exhaust your developing agents part way through.
Also if you want to go high ISO, you probably don't want to go to high dilution at the same time.
But a general methodology (just to get you going) is to shoot 3 short rolls, all the same typical scene, all at the same ISO, but bracketing exposures from -3 to +3 in one stop increments. Select a convenient dilution to mix up, and develop one roll for say 8 minutes, another for 12, and another for 16. Contact print all three strips together and examine the results. You'll hopefully find a nice diagonal to work with. It will be on your paper too, so you'll be previewing final print quality.
Best to mark the contact sheet for later reference.
The same sort of test could be done by varying the ISO instead of the development time.
Neglecting drying time, shouldn't take more than an hour to get the real world data points you need to put yourself in the right ballpark, and to know what your 'normal' prints will look like.
Just added an example, bad scan, but you get the idea. Longest development time on top, shortest on bottom.