No Code GA

this project was made for the class Evolutionary Algorithms applied to Robotics (SSC0713) at ICMC/USP (click here to see the presentation video)

This is a platform to run genetic algorithms with almost no code skills required. You can define your algorithm parameters, we'll run everything automagically and present the results in our dashboard.

Dependencies

• NodeJS
• Yarn
• MongoDB
• DynamoDB
• Serverless

How to run

• Install dependencies: run yarn

Server

• Create .env file (see .env.example);
• Run yarn server:graphql for local development

Web

• Create .env file (see .env.example);
• Run yarn web:start for local development

Architecture

• AWS Lambda Functions to run multiple algorithms and algorithm steps in parallel and distributed.
• MongoDB to persist data.
• NodeJS, React and GraphQL for the web application.

Next Steps

• Database access can be a bottleneck, find ways to improve this part
• Improve platform to support more algorithm parameters (like multi-objective optimization)

References

Applications of Evolutionary Computation