Coronavirus Optimization Algorithm: A Bioinspired Metaheuristic Based on the COVID-19 Propagation Model.

Area of Science:

• Epidemiology and Computational Intelligence
• Bioinspired Computing and Mathematical Modeling

Background:

• The rapid spread of coronavirus necessitates accurate modeling for public health interventions.
• Existing simulation models may lack flexibility in parameter setting and convergence control.

Purpose of the Study:

• To develop a novel bioinspired metaheuristic algorithm for simulating coronavirus spread.
• To integrate disease-specific parameters and dynamic factors for enhanced simulation accuracy.
• To apply the algorithm for hyperparameter optimization in deep learning and time series forecasting.

Main Methods:

• A bioinspired metaheuristic algorithm simulating coronavirus transmission dynamics.
• Incorporation of parameters such as reinfection probability, super-spreading rate, and social distancing.
• Development of a parallel multivirus version for exploring wider search spaces.
• Integration with deep learning models for hyperparameter optimization.

Main Results:

• The algorithm accurately simulates the exponential growth and subsequent decline of infected populations.
• The proposed method offers advantages in automatic parameter setting and adaptive iteration termination.
• Demonstrated remarkable performance in electricity load time series forecasting when combined with deep learning.

Conclusions:

• The novel coronavirus optimization algorithm provides an accurate and efficient tool for epidemiological modeling.
• The algorithm's adaptability and integration capabilities extend its utility to machine learning applications.
• This approach offers a promising direction for disease simulation and predictive modeling.