Optimal control of interactions between invasive alien and native species in a certain time period with the r-PINN approach
Year: 2025
Type: article
Abstract: Invasive introduced species are species that can spread to other areas and disrupt local ecosystems. Controlling the interactions between invasive species must be done in order to preserve the diversity in the ecosystem. Numerical simulations using mathematical models of reaction-diffusion systems were conducted to understand and predict how invasive species can affect river ecosystems and how optimal control can be achieved. A new approach method, r-PINN, which is a modification of the Physic-Informed Neural Network (PINN) with a proven reiteration strategy, can increase the speed of the training process. In this simulation, full control was applied in a certain time interval. When complete control was applied, the number of native species increased significantly and the number of invasive species decreased. However, the opposite was true when the control was not applied. The results of the r-PINN method used can also significantly reduce the computational time required compared to the standard PINN method. This shows that the reiteration strategy is effective in reducing computation time without losing the advantage of the loss function value.
Institutions: Universitas Ahmad Dahlan, Universiti Malaysia Terengganu
Language: English
Field: Medicine
Domain: Health Sciences
SDG: Life on land
Open Access status: gold
Estimated APC paid: $1,470
OpenAlex ID: W4409387433