A Deep Reinforcement Learning Method for Solving Task Mapping Problems with Dynamic Traffic on Parallel Systems.
YuCheng WangJerry ChouI-Hsin ChungPublished in: HPC Asia (2021)
Keyphrases
- dynamic programming
- combinatorial optimization
- cost function
- search methods
- high accuracy
- support vector machine
- objective function
- reinforcement learning
- parallel implementation
- convergence rate
- np complete
- clustering method
- segmentation method
- numerical methods
- model free
- linear systems
- quadratic programming
- distributed systems
- neural network
- significant improvement
- evolutionary algorithm
- pairwise
- search algorithm
- genetic algorithm