Predator/prey relationship need balancing
A good start, but there needs to be a way to illustrate the classic cyclical behaviour of oscillating but relatively stable populations of predator and prey. It doesn't appear possible to do that however you arrange the settings.
1. The rate at which the predators consume the herbivores and/or
2. The rate at which unfed cells die and/or
3. The rate at which cells multiply
These need careful balancing to achieve stable cyclical behaviour. What happens in the game is the predators eat until the prey population crashes totally, then the predator population crashes. It only cycles if the game values are set to have a minimum population of at least 1 cell, so the predators cannot eat the final prey, making extinction impossible. But that is a cheat in the programming to overcome a failure in the simulation. In reality, extinction is perfectly possible - but not the ONLY outcome.
I got close to a long term stable cycle by putting the plant growth at the minimum necessary for the small herbivores to survive. This meant that the predator population didn't explode, leading to immediate extinction of the herbivores. But, right on the bread-line like that, the system didn't remain stable for very long and the end result was still extinction.
If you leave out the predators, the herbivores and plants achieve a nice oscillating cyclical pattern, so well done there. Although the herbivore population rockets off the top of the graph if the plant growth is more than about 50% - should be easy enough to fix that so it doesn't happen.
Then you just need to fix the simulation for the predators, or give players control over the three aspects I mention above and make it a goal of the game to find the correct values to achieve a stable cycle.
My guess is that what you need to do is make sure predators multiply more slowly. Also, they shouldn't need to eat constantly, unlike herbivores (that's the benefit of a carnivore diet) but that might need more extensive modifications to the simulation code.