1Happ-cyber

1Happ-cyber / COHER

Public

Official implementation of the paper "Balancing Tripartite Interests in Cloud Service Composition and Optimal Selection via Curriculum-based Reinforcement Learning".

20
4
69% credibility
Found Mar 24, 2026 at 20 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A research simulation for optimizing hybrid cloud-edge-device task scheduling in manufacturing using deep reinforcement learning to balance multiple goals like time, cost, energy, and reliability.

How It Works

1
🏭 Discover Smart Factory Optimizer

You hear about a helpful tool that uses smart helpers to assign factory jobs to machines and clouds perfectly.

2
📊 Gather Your Factory Info

Collect details like job times, machine distances, and costs from your factory setup.

3
⚙️ Load Data into Simulator

Put your factory details into the ready-made playground to set the scene.

4
🤖 Train the Smart Assigners

Watch the clever brains learn to pick the best machines and clouds for every job, balancing speed, savings, and power.

5
▶️ Run Scheduling Tests

Start simulations to see jobs get assigned automatically across your factory network.

6
📈 Review Improvement Stats

Check easy charts showing less time, lower costs, better reliability, and saved energy.

🎉 Factory Runs Better!

Celebrate as your optimized schedules make everything faster, cheaper, and greener.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 20 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is COHER?

COHER is the official GitHub repository delivering a Python-based Gym environment and RL trainers for cloud-edge-device collaborative scheduling. It tackles tripartite optimization—balancing cloud providers, edge servers, and manufacturing devices—via curriculum-based reinforcement learning, minimizing makespan, energy, cost, reliability, load imbalance, and more across 200-1000 tasks. Users run training/evaluation scripts on realistic instances, generate data matrices, and test robustness to failures, ensuring coherent resource allocation without manual heuristics.

Why is it gaining traction?

It shines in multi-objective coherence where baselines falter, using curriculum stages (time+energy to full 8-objectives) and HER/OTD exploration for stable convergence on sparse rewards. Developers grab it for official GitHub actions integration, quick instance verification/generation, and ablation studies—far beyond generic RL libs lacking domain-specific envs. The coherent Deutsch-inspired tuning handles scale that coherence stream tools can't.

Who should use this?

RL practitioners optimizing hybrid cloud-edge factories with QoS tradeoffs. Cloud architects simulating device-edge-cloud flows for IoT manufacturing. Paper reproducers needing the official language implementation for coherent Göttingen/Lübeck-style benchmarks.

Verdict

Grab it for research if you're deep in DRL scheduling—20 stars and 0.7% credibility score reflect prototype maturity with thin docs, but generated instances and pre-configured TD3/DDPG agents make experiments coherent fast. Skip for production without robustness hardening.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.