EDOLAB: An Open-Source Platform for Education and Experimentation with Evolutionary Dynamic Optimization Algorithms

TL;DR

EDOLAB tackles reproducibility and accessibility challenges in evolutionary dynamic optimization by delivering an open-source, MATLAB-based platform with a dual Education and Experimentation module. It ships with 25 EDOAs, four fully parametric benchmarks (MPB, GDBG, FPs, GMPB), and two core performance indicators, $E_ ext{O}$ and $E_ ext{BBC}$, plus plots to analyze adaptation and tracking in dynamic environments. The framework emphasizes modularity and extensibility, allowing researchers to add new algorithms, benchmarks, and indicators while preserving fair comparison through standardized components and controlled randomness. EDOLAB also provides educational visualizations to help newcomers understand how EDOAs respond to environmental changes and how knowledge transfers across successive environments, with future work aimed at broader DOP coverage and a Python port.

Abstract

Many real-world optimization problems exhibit dynamic characteristics, posing significant challenges for traditional optimization techniques. Evolutionary Dynamic Optimization Algorithms (EDOAs) are designed to address these challenges effectively. However, in existing literature, the reported results for a given EDOA can vary significantly. This inconsistency often arises because the source codes for many EDOAs, which are typically complex, have not been made publicly available, leading to error-prone re-implementations. To support researchers in conducting experiments and comparing their algorithms with various EDOAs, we have developed an open-source MATLAB platform called the Evolutionary Dynamic Optimization LABoratory (EDOLAB). This platform not only facilitates research but also includes an educational module designed for instructional purposes. The education module allows users to observe: a) a 2-dimensional problem space and its morphological changes following each environmental change, b) the behaviors of individuals over time, and c) how the EDOA responds to environmental changes and tracks the moving optimum. The current version of EDOLAB features 25 EDOAs and four fully parametric benchmark generators. The MATLAB source code for EDOLAB is publicly available and can be accessed from [https://github.com/Danial-Yazdani/EDOLAB-MATLAB].

Figures (6)

• Figure 1: Examples of two-dimensional landscapes generated by the four benchmarks: (a) MPB, (b) GDBG, (c) FPs, and (d) GMPB. These are representative examples, and the characteristics shown cannot be generalized to all possible landscapes generated by these benchmarks.
• Figure A2-1: A general sequence diagram of running an EDOA in EDOLAB.
• Figure A2-2: The experimentation module of EDOLAB.
• Figure A2-3: An Excel output table generated by EDOLAB's $\mathtt{OutputExcel.m}$ function.
• Figure A2-4: An output figure of an experimentation in EDOLAB. This figure depicts the plots of offline and current errors over time. The plots are the average of all runs.