It’s certainly been a while. It hasn’t felt like it though. So much has happened.
I got hired at Gree International Entertainment. Worked there for 18 months until they decided to shut down all international operations, meaning the immediate redundancy of around 200ish people across Melbourne, San Francisco, and Berlin. We were working on a mobile game that was based off a shitton of consumer insight testing. They’d spent months figuring out that people were interested in zombie apocalypses with evolution themes. Go figure.
This experience renewed my faith in Strain. I’ve been thinking about it much more as of late. I’m considering whether focusing on the evolution of the zombies is the right path – particularly if art ends up being an issue. Perhaps the better option is to focus instead on AI behaviour and trying to create emergent gameplay by creating a simple set of rules that govern all of the characters: zombie and human alike.
This train of thought has today led me down a rabbit hole of scholarly articles and broken links. Agent based modelling is something I dipped my toes in while studying. Netlogo served as our tool. It’s a fairly robust piece of software that lets you run your own agent based simulations in which you set up the rules of interaction. I’m wanting to create my own agent-based modelling solution in Unity (or UE4?) that can simulate large numbers of actors (I’m thinking in the order of the hundreds of thousands right now). To get this running, I’m planning on using Compute Shaders to do as much of the heavy lifting on the GPU as possible. I’ve found a few papers that go into agent based modelling approaches that take advantage of GPGPU but have yet to read them.
I’m anticipating issues concerning how to get agents properly interacting with one-another given the highly parallelised nature of GPGPU. How can one agent get info from every other agent around it if, in its own thread, it only has visibility of itself? Perhaps using some kind of vector field or other environment state modifications to determine how they move? Hopefully reading these papers will give me some more insight about how to achieve all of this.
I also created an itch.io page because why not? Unsure whether I’ll put the rest of the dev log over there or keep it here. Also unsure what I’m going to do in terms of art in the medium to long term. That’s a problem for future Andrew to worry about.