.png){kind=icon}
.png){kind=icon}
Return
State prediction for multiple diffusion targets based on point pattern physics-informed neural network
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.129714.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
State prediction for multiple diffuse targets focuses on locating diffuse sources and concentration distributions. However, in real scenarios, problems such as multi-target overlap, unknown diffusion parameters, and difficult- to-measure environmental disturbances arise, causing the existing algorithms to have poor prediction accuracy. To address the above problems, this article proposes a multi-diffusion target state prediction algorithm based on point pattern physics-informed neural network (PP-PINN). This article targets the overlapping problem of multiple diffusion targets and uses the point pattern model to realize the accurate clustering of different diffusion targets and estimate their concentration distributions. Then, the physics-informed neural network is trained to estimate the diffusion parameters and modify the initial concentration distribution to reduce the environmental interference. The experimental results show that the proposed algorithm can significantly improve the performance of multi-diffusion target state prediction, and can provide important data support for the assessment of hazardous material leakage accidents.
Keywords:
State prediction
Multiple diffusion targets
Concentration distribution
Physics-informed neural network
Point pattern
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W