Please wait ...
0% Complete
فارسی
Home
/
شانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش
TDO-SA-PINN: A Co-Evolutionary Framework for Physics-Informed Neural Networks
Authors :
SeyedMohammadReza AhmadEnjavi
1
Masoud Shafiee
2
1- دانشگاه صنعتی امیرکبیر
2- دانشگاه صنعتی امیرکبیر
Keywords :
Physics-Informed Neural Networks،Tasmanian Devil Optimizer،Optimization for Deep Learning
Abstract :
Physics-Informed Neural Networks (PINNs) have emerged as a promising paradigm for solving forward and inverse partial differential equations (PDEs), yet their performance often deteriorates in stiff, multi-scale, or high-frequency regimes due to spectral bias, loss imbalance, and local optimization pathologies. While Self-Adaptive PINNs (SA-PINNs) mitigate error concen tration by dynamically adjusting residual weights, their correc tive power remains constrained by gradient-based optimizers that stagnate in rugged landscapes. To address this gap, we introduce a co-evolutionary framework that integrates SA-PINNs with the Tasmanian Devil Optimizer (TDO), a recent population based metaheuristic. In the proposed TDO-SA-PINN, adaptive weights reshape the loss landscape while a diverse swarm of candidate networks performs global, gradient-free exploration. This dual mechanism simultaneously targets spectral bias and optimizer-induced stagnation, and naturally yields an ensemble that encodes predictive uncertainty. Extensive experiments on canonical PDE benchmarks demonstrate that TDO-SA-PINNs achieve lower error and more reliable convergence compared to standard PINNs trained with ADAM/LBFGS, adaptive PINN variants, and deep ensembles. The results highlight the potential of co-evolutionary population search as a scalable and effective complement to adaptive physics-informed learning frameworks.
Papers List
List of archived papers
Designation and development of Camera Sensor for identification of polymer type
Negin Piri - Ahmad Salehi - Erfan Memarian
Combinatorial Auction Based on Social Choice in the Internet of Things
Maede Esmaeili - Faria Nassiri-Mofakham - Fatemeh Hassanvand
Real-Time EEG-Based Analysis Of Stress-Inducing Stimuli
Mohsen Mahmoudi - Fattaneh Taghiyareh - Yasamin Akhavein - Elnaz Ghorbani
Enhancing Employee Promotion Prediction with a Novel Hybrid Model Integrating Convolutional Neural Networks and Random Forest
Pouya Ardehkhani - Seyyed Reza Moslemi - Hanieh Hooshmand
Distributed Learning Automata-based Algorithm for Finding K-Clique in Complex Social Networks
Mohammad Mehdi Daliri Khomami - Alireza Rezvanian - Ali Mohammad Saghiri - Mohammad Reza Meybodi
Benchmarking Embedding Models for Persian-Language Semantic Information Retrieval
Mahmood Kalantari - Mehdi Feghhi - Nasser Mozayani
ISPREC: Integrated Scientific Paper Recommendation using heterogeneous information network
Elaheh Jafari - Dr Bita Shams - Dr Saman Haratizadeh
Using Trust Statements and Ratings by GraphSAGE to Alleviate Cold Start in Recommender Systems
Seyedeh Niusha Motevallian - Dr Seyed Mohammad Hossein Hasheminejad
Detection and Identification of Cyber-Attacks in Cyber-Physical Systems Based on Machine Learning Methods
Zohre Nasiri Zarandi
Sustainability analysis and improvement of model driven engineering and model transformation languages
Kevin Lano - Shekoufeh Kolahdouz Rahimi
more
Samin Hamayesh - Version 45.0.1