Deadlock situation
Deadlock -free graph

ps: the demo video for this project is available at github , thank you

Inspiration I found os a difficult subject in college and when I searching for educational ideas , i felt this would be good idea to handle.

What it does Helps students learn os concepts and faculties can teach custom deadlocks to students in a more interactive way.Think of it as a sandbox where you can watch an AI actually "use a computer" instead of just answering questions about one.

How I built it I stitched together a simulated desktop environment with a vision-and-action loop, A lot of the glue code was just making sure the feedback loop between "AI sees state" and "AI acts on state" stayed fast and reliable.

Challenges I ran into TypeScript and React flow was new to me , learning it in a short time was challenging as well as rewarding . I also burned a lot of time on timing issues like how the AI would act before the UI had actually finished loading, causing it to click on things that weren't there yet.

Accomplishments that I'm proud of We're proud we actually got a closed loop working end to end perception, decision, action — without it falling over constantly. Seeing the AI complete a multi-step task on its own for the first time felt like a genuine "it's alive" moment for the team.

What we learned We learned that simulating an environment is often harder than building the intelligence that operates in it. We also came away with a much deeper respect for how much humans take for granted when navigating a screen.

What's next for OS AI Simulation Project I would like to add more simulations like cpu scheduling , banker's algo , etc. and maybe even deploy if everything goes good