MEALPY: An open-source library for latest meta-heuristic algorithms in Python

Abstract

Meta-heuristic algorithms are becoming more prevalent and have been widely applied in various fields. There are numerous reasons for the success of such techniques in both science and industry, including but not limited to simplicity in search/optimization mechanisms, implementation readiness, black-box nature, and ease of use. Although the solutions obtained by such algorithms are not guaranteed to be exactly global optimal, they usually find reasonably good solutions in a reasonable time. Many algorithms have been proposed and developed in the last two decades. However, there is no library implementing meta-heuristic algorithms, which is easy to use and has a vast collection of algorithms. This paper proposes an open-source and cross-platform Python library for nature-inspired optimization algorithms called Mealpy. To propose Mealpy, we analyze the features of existing libraries for meta-heuristic algorithms. After, we propose the designation and the structure of Mealpy and validate it with a case study discussion. Compared with other libraries, our proposed Mealpy has the largest number of classical and state-of-the-art meta-heuristic algorithms, with more than 160 algorithms. Mealpy is an open-source library with well-documented code, has a simple interface, and benefits from minimum dependencies. Mealpy includes a wide range of well-known and recent meta-heuristics algorithms capable of optimizing challenge benchmark functions (e.g. CEC-2017). Mealpy can also be used for practical problems such as optimizing parameters for machine learning models. We invite the research community for widespread evaluations of this comprehensive library as a promising tool for research study and real-world optimization. The source codes, supplementary materials, and guidance is publicly available on GitHub: https://github.com/thieu1995/mealpy.