You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
🌟 TensorACO: Tensorized Ant Colony Optimization for GPU Acceleration 🌟
Tensorized Ant Colony Optimization (TensorACO) enhances the convergence speed and efficient of large-scale Traveling Salesman Problems (TSP) by incorporating GPU acceleration. By tensorizing the ant system and path, TensorACO capitalizes on GPU parallelism for accelerated computation. Additionally, the Adaptive Independent Roulette (AdaIR) method enhances the performance by a dynamically strategy. TensorACO is compatible with the EvoX framework.
Key Features
GPU Acceleration 💻: Leverages GPUs for enhanced computational capabilities.
Large-Scale Optimization 📈: Ideal for large ant colony sizes and large city sizes.
Real-World Applications 🌐: Suited for TSP, a central challenge in combinatorial optimization, is characterized by the theoretical complexity and practical relevance in routing and logistics.
If you use TensorACO in your research and want to cite it in your work, please use:
@article{tensoraco,
title = {{Tensorized} {Ant} {Colony} {Optimization} {for} {GPU} {Acceleration}},
author = {Yang, Luming and Jiang, Tao and Cheng, Ran},
booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference (GECCO)},
year = {2024}
}