Plus the weights in your neural network are just an array of numbers, very ripe for mutations and other evolution-like algorithms. You will usually end up with something that spins in circles, doesn't move, runs from prey and towards predators, but those genomes don't survive to the next generation. Have you considered a neural net? The weights of the nodes in the network can be part of your 'genome', inputs will be direction and distance to prey/predator, current energy, current speed and direction, etc. Perhaps tentacles represented by a series of chained spheres? How can I add physical growth to these organisms is my overarching question. I guess it's difficult to implement something beyond mere size, but I guess that's the question I'm asking. The generations would reach an optimal size. Smaller ones would maneuver better, obviously. In the sense that the Darwin floaters in the linked example are maneuvering physically through water.is there any sort of way to evolve organisms that are capable of interacting with an environment? Hypothetically, I create obstacles in the environment. I think I'm trying to figure out how to develop organisms that aren't simply circles with vaguely hidden genetic traits and abilities. Any thoughts would be appreciated.ĮDIT: I guess I should clarify my question a little. But how could I also implement things like advanced behaviors? For example, maybe moving slowly towards prey before speeding? I've considered that perhaps their acceleration towards prey is determined by a polynomial with genetically-defined coefficients. This site has a great motility evolution simulation. How can I implement complexity? Multicellular? How about motility, mouths? A Spore-like design? But it wouldnt' be too interesting, either. My basic broad question is as thus: doing this with unicellular blobs wouldn't be TOO hard. There will be a pretty simple AI loop where the thresholds of each activity are based on their genetic variables. All organisms will have a genetic sequence consisting of traits such as mass, speed, vision range, metabolism, how close a predator gets before running, how close prey gets before running, how low energy gets before stopping current activities in favor of hunger, etc, etc. I'm brainstorming a simulation game that should ideally mimic evolution.

Darwin pond game