ddmtolab 1.0.0

Data-Driven Multitask Optimization Laboratory

DDMTOLab

Data-Driven Multitask Optimization Laboratory

---

📖 Overview

DDMTOLab is a comprehensive Python platform designed for data-driven multitask optimization, featuring 30+ algorithms, 100+ benchmark problems, and powerful experiment tools for algorithm development and performance evaluation.

Whether you're working on expensive black-box optimization, multi-objective optimization, or complex multi-task scenarios, DDMTOLab provides a flexible and extensible framework to accelerate your research and support real-world applications.

✨ Features

• 🚀 Comprehensive Algorithms - Single/multi-task, single/multi-objective optimization algorithms
• 📊 Rich Problem Suite - Extensive benchmark functions and real-world applications
• 🤖 Data-Driven Optimization - Surrogate modelling for expensive optimization
• 🔧 Flexible Framework - Simple API and intuitive workflow for rapid prototyping
• 🔌 Fully Extensible - Easy to add custom algorithms and problems
• 📈 Powerful Analysis Tools - Built-in visualization and statistical analysis
• ⚡ Parallel Computing - Multi-core support for batch experiments
• 📝 Complete Documentation - Comprehensive Tutorial and API reference

🚀 Quick Start

👉 Get Started with Our Tutorial

DDMTOLab requires:

• Python 3.10+
• PyTorch 2.5+ with CUDA 12.1 support (optional, for GPU-accelerated Gaussian Process modeling)
• BoTorch 0.16+
• GPyTorch 1.14+
• NumPy 2.0+
• SciPy 1.15+
• scikit-learn 1.7+
• Pandas 2.3+
• Matplotlib 3.10+
• tqdm

Installation

git clone https://github.com/JiangtaoShen/DDMTOLab.git
cd DDMTOLab
pip install -r requirements.txt

Basic Usage

import numpy as np
from ddmtolab.Methods.mtop import MTOP
from ddmtolab.Algorithms.STSO.GA import GA

# Step 1: Define objective function
def t1(x):
    """Forrester function: (6x-2)^2 * sin(12x-4)"""
    return (6 * x - 2) ** 2 * np.sin(12 * x - 4)

# Step 2: Create optimization problem
problem = MTOP()
problem.add_task(t1, dim=1)

# Step 3: Run optimization
results = GA(problem).optimize()

# Step 4: Display results
print(results.best_decs, results.best_objs)

# Step 5: Analyze and visualize
from ddmtolab.Methods.test_data_analysis import TestDataAnalyzer
TestDataAnalyzer().run()

Batch Experiments

from ddmtolab.Methods.batch_experiment import BatchExperiment
from ddmtolab.Methods.data_analysis import DataAnalyzer
from ddmtolab.Algorithms.STSO.GA import GA
from ddmtolab.Algorithms.STSO.PSO import PSO
from ddmtolab.Algorithms.STSO.DE import DE
from ddmtolab.Algorithms.MTSO.EMEA import EMEA
from ddmtolab.Algorithms.MTSO.MFEA import MFEA
from ddmtolab.Problems.MTSO.cec17_mtso import CEC17MTSO

if __name__ == '__main__':
    # Step 1: Create batch experiment manager
    batch_exp = BatchExperiment(
        base_path='./Data',      # Data save path
        clear_folder=True        # Clear existing data
    )

    # Step 2: Add test problems
    cec17mtso = CEC17MTSO()
    batch_exp.add_problem(cec17mtso.P1, 'P1')
    batch_exp.add_problem(cec17mtso.P2, 'P2')
    batch_exp.add_problem(cec17mtso.P3, 'P3')

    # Step 3: Add algorithms with parameters
    batch_exp.add_algorithm(GA, 'GA', n=100, max_nfes=20000)
    batch_exp.add_algorithm(DE, 'DE', n=100, max_nfes=20000)
    batch_exp.add_algorithm(PSO, 'PSO', n=100, max_nfes=20000)
    batch_exp.add_algorithm(MFEA, 'MFEA', n=100, max_nfes=20000)
    batch_exp.add_algorithm(EMEA, 'EMEA', n=100, max_nfes=20000)

    # Step 4: Run batch experiments
    batch_exp.run(n_runs=20, verbose=True, max_workers=8)

    # Step 5: Configure data analyzer
    analyzer = DataAnalyzer(algorithm_order=['GA', 'DE', 'PSO', 'EMEA', 'MFEA'])

    # Step 6: Run data analysis
    results = analyzer.run()

📊 Example Results

Results from the batch experiment above:

🎯 Key Components

Algorithms

40+ state-of-the-art optimization algorithms across four categories:

Category	Algorithms
STSO	GA, DE, PSO, SL-PSO, KL-PSO, CSO, CMA-ES, AO, GWO, EO, BO, EEI-BO
STMO	NSGA-II, NSGA-III, NSGA-II-SDR, SPEA2, MOEA/D, MOEA/DD, RVEA, IBEA, Two_Arch2, MSEA, CCMO
MTSO	MFEA, MFEA-II, EMEA, G-MFEA, MTBO, RAMTEA, SELF, EEI-BO+, MUMBO, LCB-EMT
MTMO	MO-MFEA, MO-MFEA-II

Problems

132+ benchmark problems across five categories:

Category	Problem Suites
STSO	Classical Functions (9), CEC10-CSO (18)
STMO	ZDT (6), UF (10), DTLZ (9)
MTSO	CEC17-MTSO (9), CEC17-MTSO-10D (9), CEC19-MaTSO (6)
MTMO	CEC17-MTMO (9), CEC19-MTMO (10), CEC19-MaTMO (6), CEC21-MTMO (10), MTMO-Instance (2)
Real-World	PEPVM (1), PINN-HPO (12), SOPM (2), SCP (1), MO-SCP (2), PKACP (1)

Methods

• Batch Experiments: Parallel execution framework for large-scale testing
• Data Analysis: Statistical testing (Wilcoxon, Friedman) and visualization tools
• Performance Metrics: IGD, HV, Spacing, Spread, FR, CV, and more
• Algorithm Components: Reusable building blocks for rapid development

📄 Citation

If you use DDMTOLab in your research, please cite:

@software{ddmtolab2025,
  author = {Jiangtao Shen},
  title = {DDMTOLab: A Python Platform for Data-Driven Multitask Optimization},
  year = {2025},
  url = {https://github.com/JiangtaoShen/DDMTOLab}
}

📧 Contact

• Author: Jiangtao Shen
• Email: [e-mail]
• Documentation: https://jiangtaoshen.github.io/DDMTOLab/
• Issues: GitHub Issues

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Made with ❤️ by Jiangtao Shen