Variable Problem Dimensionality in Evolutionary Computation (VarDimInEC)

It’s been a long time since MessyGA.  

Description

In the realm of AI, addressing complex optimization challenges with variable problem dimensionality remains an underexplored area. Existing meta-heuristic algorithms excel in solving optimization problems but often struggle when faced with problems of changing dimensionality. By addressing this critical research gap, the research findings not only advance the field of AI-driven optimization but also offer practical solutions to enhance the performance and applicability of meta-heuristic algorithms in dynamic and uncertain environments. The results and methodology may have significant implications for a wide range of domains, such as network design, resource allocation, and machine learning, where variable problem dimensionality is prevalent and challenging. 

This special session aims to bridge this gap by discussing the original research focused on adaptive strategies for meta-heuristic optimization tailored to variable problem dimensionality. Such research may become a vitally important part of science and engineering at the theoretical and practical levels. The research papers presenting the extensive analysis of the proposed methods, using both benchmark problems and real-world applications, as well as discussing the research focusing on evaluating the adaptability of these algorithms, comparing their performance, and assessing the trade-offs between algorithm complexity and adaptability, are welcomed. 

Our aim with this special session is to bring together experts from fundamental research of evolutionary computation and experts from various evolutionary computation application fields researching and developing novel algorithmic approaches capable of dynamically adjusting search operators, population sizes, and exploration-exploitation strategies in response to variations in problem dimensionality. The scope will be focused on, but not limited to, the following topics.

List of topics

• Applications of nature-inspired computing and optimization algorithms for problems with variable dimensionality. 
• The performance improvement, benchmarking, testing and efficiency of the proposed algorithms. 
• Creation of new benchmarking sets. 
• Adaptive strategies for meta-heuristic optimization tailored to variable problem dimensionality. 
• Visualizations and methodology for analysis of population dynamics with variable dimensionality problems. 
• Adjusting state-of-the-art meta-heuristic techniques for global optimization to the problems with variable dimensionality. 

Organizers