Towards Complex Systems Behavioral Prediction: A Survey of Artificial Intelligence Applications

Because of the subsystem interconnections, dependencies, and unknown behavior of these systems, modeling and prediction of complex systems is a difficult challenge. These traits present a significant barrier in precisely anticipating the states of development of complex systems. Most naturally occurring systems, from the human body to the environment, may be classified as complex due to high degrees of interconnectivity, dependency among components, and flexible behaviors. The most frequent methods used to study and forecast the behavior and performance of these complex systems are mathematical modeling and computational intelligence. The current research investigates the advances achieved in complex system prediction utilizing artificial intelligence (AI) based 2approaches to date. A total of 18 scientific research publications describing the application of AI-based approaches in complex system prediction were chosen from various academic sources. These research encompass a variety of AI sapproaches such as machine learning and deep learning algorithms. This work presents an assessment of the available strategies and a comparative analysis of their performances while covering several existing AI-based techniques in complex systems prediction. The performance comparison was presented in terms of the datasets and assessment measures that were employed.