b1f6c1c4 / ansys-moe

Computer-automated Design (CautoD) System 自动化设计系统

ansys-moe 自动化设计系统

Computer-automated Design (CautoD) System

Introduction

Using Bayesian Global Optimization (BGO) with LHSMDU and EPI, this system automates the solution of optimal design problems, specifically those involving Ansys, Python, R, and Mathematica. See thesis (in Chinese) folder for more detail.

Project structure

Folder	Tech Stack	Deployment	License
thesis 论文	LaTeX	n/a	Proprietary / GNU AGPLv3
commond 计算服务	Golang	docker/binary	GNU AGPLv3
controller 工作流内核	JavaScript	docker	GNU AGPLv3
facade 网站后端	JavaScript	docker	GNU AGPLv3
frontend 网站前端	React+Redux	docker	MIT
nginx 反向代理	nginx	docker	GNU AGPLv3
storage 文件存储	JavaScript	docker	GNU AGPLv3
elastalert 告警	ELK	docker	GNU AGPLv3
filebeat 日志上传	ELK	docker	GNU AGPLv3
logstash 日志处理	ELK	docker	GNU AGPLv3

Build

thesis:

git submodule init
make

commond, controller, facade, storage:

vim .env # Set build machine and dockerhub credential
make vendor
make

frontend:

eval $(docker-machine env <SRC_BUILD_MACHINE>)
yarn build

nginx, elastalert, filebeat, logstash:

docker-compose build

Notice: build frontend before build nginx

Run

docker-compose up

Notice: data stored in /mnt/ansys-moe

Notice: one should run artifect commond-std or commond-svc from commond to enable Ansys and Mathematica support.

Legal Stuff

commond, controller, elastalert, facade, filebeat, frontend, logstash, nginx, and storage components are licensed under GNU AGPLv3 License, frontend component is licensed under MIT License, while thesis is Proprietary work with all right reserved (except thesis/data, which is licensed under GNU AGPLv3 License).

Besides, the author has distributed this project to Tsinghua University as well, under different licenses, and authorized Tsinghua University to store and publish thesis.