Mixture of Experts for Network Optimization: A Large Language Model-enabled Approach

Hongyang Du, Guangyuan Liu, Yijing Lin, Dusit Niyato, Jiawen Kang, Zehui Xiong, Dong In Kim
Nanyang Technological University
Paper Code arXiv

Abstract

Optimizing various wireless user tasks poses a significant challenge for networking systems because of the expanding range of user requirements. Despite advancements in Deep Reinforcement Learning (DRL), the need for customized optimization tasks for individual users complicates developing and applying numerous DRL models, leading to substantial computation resource and energy consumption and can lead to inconsistent outcomes. To address this issue, we propose a novel approach utilizing a Mixture of Experts (MoE) framework, augmented with Large Language Models (LLMs), to analyze user objectives and constraints effectively, select specialized DRL experts, and weigh each decision from the participating experts. Specifically, we develop a gate network to oversee the expert models, allowing a collective of experts to tackle a wide array of new tasks. Furthermore, we innovatively substitute the traditional gate network with an LLM, leveraging its advanced reasoning capabilities to manage expert model selection for joint decisions. Our proposed method reduces the need to train new DRL models for each unique optimization problem, decreasing energy consumption and AI model implementation costs. The LLM-enabled MoE approach is validated through a general maze navigation task and a specific network service provider utility maximization task, demonstrating its effectiveness and practical applicability in optimizing complex networking systems.

The following figure shows two network optimization strategies. Part A demonstrates the drawbacks of training distinct AI models for different user requirements, emphasizing the costs of excessive AI model deployment. Part B presents our LLM-enabled MoE approach, using a limited set of DRL models to efficiently address a variety of user tasks.

Network Optimization Strategies

The problem we solve: How can we achieve effective network optimization without using numerous DRL models individually trained for each specific task?

🔧 Environment Setup

To establish a new conda environment, run the following command:

conda create --name moeopt python==3.7

⚡ Activate Environment

Activate the environment with:

conda activate moeopt

📦 Install Required Packages

Install the necessary packages one by one using pip:

pip install torch
pip install opencv-python==4.1.2.30
pip install scipy
pip install torchvision
pip install scikit-image

🏃‍♀️ Run the Program

Execute main.py to initiate the program.

🔍 Case Study 1: Maze Navigation Task

LLM-enabled MoE framework demonstrated in a maze navigation task, using an ensemble of DRL models for diverse tasks as expert models for the LLM to infer and address user tasks.

Maze Navigation Task

🔍 Case Study 2: Service Provider Utility Maximization Problem

Illustration of the system model, market interactions, and the impact of varying user requirements on service payments and power allocation strategies.

Service Provider Utility Maximization Problem

📚 Cite Our Work

If our code aids your research, please cite our work:

@article{du2024mixture,
  title={Mixture of Experts for Network Optimization: A Large Language Model-enabled Approach},
  author={Du, Hongyang and Liu, Guangyuan and Lin, Yijing and Niyato, Dusit and Kang, Jiawen and Xiong, Zehui and Kim, Dong In},
  journal={arXiv preprint arXiv:2402.09756},
  year={2024},
  month = {Feb.}
}