Optima-CityU/LLM4AD_Next

A next-generation automatic algorithm design platform, making automated algorithm design more accessible and easier to use

What it solves

LLM4AD_Next removes the high entry barrier and tedious manual configuration typically required when using Large Language Models for Automated Algorithm Design (LLM4AD). It automates the transition from a natural-language problem description to a fully runnable evolutionary algorithm search pipeline.

How it works

The system uses an interactive AI consultant (llm4ad chat) to interview the user and understand their requirements. It then automatically generates the necessary framework components, including an evaluator, algorithm skeleton, configuration, and debugger. To optimize the algorithms, it combines LLM-powered design with evolutionary optimization methods (such as EoH, MEoH, ReEvo, and MCTS-AHD) to evolve high-performing code. It also includes an Evolve-Block Advisor that can scan existing repositories to recommend specific code blocks for evolution.

Who it’s for

It is designed for researchers and developers who want to automate the discovery and optimization of algorithms without manually setting up complex configuration pipelines.

Highlights

  • Interactive Setup: A conversational AI consultant that generates a ready-to-run app framework from a problem description.
  • Evolutionary Search: Supports multiple automated heuristic design methods including IslandGA, DyCA, EoH, MEoH, ReEvo, and MCTS-AHD.
  • Code Analysis: An Evolve-Block Advisor that identifies and scores promising code blocks in any repository for further evolution.
  • Long-term Memory: Integration with MindMemOS for global, project, and task-scoped memory.

Related

  • Project
    SakanaAI/ShinkaEvolve

    A framework that combines LLMs and evolutionary algorithms to automate the discovery and optimization of scientific code across various programming languages.

  • Project
    algorithmicsuperintelligence/openevolve

    An open-source evolutionary coding agent that uses LLMs and MAP-Elites to autonomously discover and optimize algorithms and hardware-specific code.

  • Project
    AMA-CMFAI/LAMBDA

    A data analysis agent that converts natural-language questions into reproducible analysis workflows, including code execution, visualization, and report generation.

  • Project
    ANative-Lab/EvoAgentX

    An open-source framework for automatically building, evaluating, and evolving LLM-based agentic workflows through iterative feedback loops.

  • Project
    microsoft/LMOps

    A research initiative focused on fundamental technologies to improve LLM prompting, inference speed, and context handling for building AI products.