PlatEMO: A MATLAB Platform for Evolutionary Multi-Objective Optimization (1701.00879v1)

Abstract: Over the last three decades, a large number of evolutionary algorithms have been developed for solving multiobjective optimization problems. However, there lacks an up-to-date and comprehensive software platform for researchers to properly benchmark existing algorithms and for practitioners to apply selected algorithms to solve their real-world problems. The demand of such a common tool becomes even more urgent, when the source code of many proposed algorithms has not been made publicly available. To address these issues, we have developed a MATLAB platform for evolutionary multi-objective optimization in this paper, called PlatEMO, which includes more than 50 multi-objective evolutionary algorithms and more than 100 multi-objective test problems, along with several widely used performance indicators. With a user-friendly graphical user interface, PlatEMO enables users to easily compare several evolutionary algorithms at one time and collect statistical results in Excel or LaTeX files. More importantly, PlatEMO is completely open source, such that users are able to develop new algorithms on the basis of it. This paper introduces the main features of PlatEMO and illustrates how to use it for performing comparative experiments, embedding new algorithms, creating new test problems, and developing performance indicators. Source code of PlatEMO is now available at: http://bimk.ahu.edu.cn/index.php?s=/Index/Software/index.html.

• The paper presents a MATLAB-based open-source platform that integrates 50+ evolutionary algorithms and 110 benchmark problems to standardize multi-objective optimization research.
• It provides an intuitive GUI and efficient MATLAB coding capabilities to enable seamless benchmarking and extension of optimization algorithms.
• The platform’s robust design enhances reproducibility and accelerates the development of innovative evolutionary multi-objective optimization methods.

PlatEMO#1: A MATLAB Platform for Evolutionary Multi-Objective Optimization

The paper "PlatEMO#1: A MATLAB Platform for Evolutionary Multi-Objective Optimization" authored by Ye Tian, Ran Cheng, Xingyi Zhang, and Yaochu Jin presents a comprehensive MATLAB-based platform engineered to facilitate evolutionary multi-objective optimization (EMO) research and applications. This initiative addresses critical challenges in the field by offering a robust, open-source software environment that encompasses a wide variety of evolutionary algorithms and optimization problems.

Overview and Features

PlatEMO#1 integrates over 50 multi-objective evolutionary algorithms (MOEAs) and more than 100 multi-objective test problems. This extensive repository offers practitioners and researchers a robust framework to benchmark existing algorithms and develop new ones. Key features of PlatEMO#1 include:

1. Rich Library: The platform includes numerous state-of-the-art MOEAs published in top-tier journals, covering a broad spectrum of algorithms such as multi-objective genetic algorithms (MOGAs), multi-objective differential evolution algorithms, multi-objective particle swarm optimization algorithms, and others.
2. Good Usability: PlatEMO#1 is developed entirely in MATLAB, ensuring compatibility across various operating systems. It leverages MATLAB's matrix operations to ensure concise and efficient coding. Additionally, the platform offers a highly intuitive graphical user interface (GUI) that allows users to execute complex optimization tasks without exploring code.
3. Easy Extensibility: The open-source nature of PlatEMO#1 facilitates the extension and customization of the platform. Researchers can integrate new algorithms, optimization problems, operators, and performance indicators by following the provided interfaces and guidelines.
4. Delicate Considerations: The platform incorporates several nuanced features designed to enhance functionality and performance, such as efficient non-dominated sorting algorithms and support for various Pareto front shapes.

Numerical Results and Bold Claims

The numerical robustness of PlatEMO#1 is underscored by its inclusion of 50 contemporary MOEAs and 110 benchmark multi-objective problems, each implemented succinctly thanks to MATLAB’s efficient matrix operations. Table 1 and Table 2 enumerated in the paper detail the included algorithms and test problems respectively. Among these, notable algorithms such as NSGA-II, SPEA2, and MOEA/D are showcased, alongside many-objective optimizers like RVEA and NSGA-III.

Implications and Future Developments

The practical implications of PlatEMO#1 are significant, as it provides both a developmental and comparative framework for EMO researchers. The ease of implementing and benchmarking new algorithms can accelerate advancements in the field of multi-objective optimization. Moreover, users can readily convert experimental results into formats suitable for academic presentation, such as Excel and LaTeX tables.

Theoretically, PlatEMO#1’s contribution lies in standardizing the experimental environment, thereby ensuring reproducibility and consistency in multi-objective optimization research. This may lead to more rigorous empirical studies and foster a deeper understanding of algorithmic performance across diverse problem sets.

Future developments for PlatEMO#1 may include the incorporation of modules for niche optimization tasks such as preference optimization, dynamic optimization, and noisy optimization. Additionally, expanding the repository of algorithms and test problems to encompass the latest innovations in the field will maintain PlatEMO#1's relevance and utility.

In summary, PlatEMO#1 stands as a valuable resource in the evolutionary computation community, addressing the dual need for a comprehensive benchmarking tool and a flexible development platform. Its user-centric design, coupled with an extensive array of algorithms and test problems, posits it as a key asset for researchers and practitioners focusing on multi-objective optimization.