Evolutionary Dynamic Optimization and Machine Learning
Abdennour Boulesnane
TL;DR
The paper surveys Evolutionary Dynamic Optimization (EDO) and its reciprocal relationship with Machine Learning (ML), arguing that ML can enhance dynamic optimization by leveraging data from iterative evolutionary searches, while EC methods can optimize ML tasks subject to dynamic objectives. It reviews ML paradigms—Transfer Learning, Supervised, Reinforcement, and Unsupervised learning—and their application to Dynamic Optimization Problems (DOPs), including algorithms such as RL-DMOEA, MOEA/D-GMM, and Q-learning-based approaches. It highlights the use of EDO in ML tasks, notably streaming feature selection, dynamic hyperparameter tuning, and neural architecture search, with examples like WD2O/OSFS, SMiLE, and CNN hyperparameter Q-learning. The authors argue this reciprocal integration can improve convergence, robustness, and adaptability in dynamic, real-world problems and call for further research into novel EDO-ML hybrids. Overall, the work provides a first comprehensive overview and sets a roadmap for advancing evolutionary learning in dynamic ML domains.
Abstract
Evolutionary Computation (EC) has emerged as a powerful field of Artificial Intelligence, inspired by nature's mechanisms of gradual development. However, EC approaches often face challenges such as stagnation, diversity loss, computational complexity, population initialization, and premature convergence. To overcome these limitations, researchers have integrated learning algorithms with evolutionary techniques. This integration harnesses the valuable data generated by EC algorithms during iterative searches, providing insights into the search space and population dynamics. Similarly, the relationship between evolutionary algorithms and Machine Learning (ML) is reciprocal, as EC methods offer exceptional opportunities for optimizing complex ML tasks characterized by noisy, inaccurate, and dynamic objective functions. These hybrid techniques, known as Evolutionary Machine Learning (EML), have been applied at various stages of the ML process. EC techniques play a vital role in tasks such as data balancing, feature selection, and model training optimization. Moreover, ML tasks often require dynamic optimization, for which Evolutionary Dynamic Optimization (EDO) is valuable. This paper presents the first comprehensive exploration of reciprocal integration between EDO and ML. The study aims to stimulate interest in the evolutionary learning community and inspire innovative contributions in this domain.