KKT-based trade-off explanations for interactive multiobjective optimization

This repository contains the implementation of software used in the research article "Can LIME Make Interactive Multiobjective Optimization Methods Explainable?"

The repository includes two main components:

• Backend: Computes KKT multipliers and approximated solutions. It provides a Flask API for integration with the frontend but can also function as a standalone module.
• Frontend: An interactive user interface that allows users to input reference points and obtain non-dominated solutions with explanations.

Installation

To get started, clone this repository:

git clone https://github.com/industrial-optimization-group/kkt-lime-explanations.git

Backend Setup

1. Create a virtual environment for the backend Using Python's venv module:

python -m venv .venv

Or with Poetry:

poetry shell

2. Activate the virtual environment For venv:

source .venv/bin/activate

Or with Poetry:

poetry shell

3. Install dependencies Navigate to the tradeoff-analysis folder and install the backend code:

poetry install

4. Run the code Navigate to the explainable_moo folder and run:

python api.py

After this, the server should be running locally in your computer. The IP address of the server will be shown in the terminal.

Installing the frontend

1. In a separate terminal, navigate to the XMOO-UI folder:

cd kkt-lime-explanations
cd XMOO-UI

2. Install Dependencies: Use npm, pnpm, or yarn to install the required packages:

npm install

(Replace npm with pnpm or yarn if using one of those package managers.)

3. Run the development server: Start the server for local development:

npm run dev

Or open the app in a browser automatically:

npm run dev -- --open

Usage

1. Ensure the backend server is running.

• Activate the virtual environment and start the Flask API server from the backend directory.

2. Access the frontend application:

• Open the development server URL (default: http://localhost:3000) in a web browser.

3. Provide reference points in the frontend to generate solutions and explanations.