A Reinforcement Learning Framework for Combinatorial Optimization

Proceedings of the AAAI Conference on Artificial Intelligence, 13

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 13

AAAI-96 Student Abstracts

The combination of reinforcement learning methods with neural networks has found success on a growing number of large-scale applications, including backgammon move selection, elevator control, and job-shop scheduling. In this work, we modify and generalize the scheduling paradigm used by Zhang and Dietterich to produce a general reinforcement-learning-based framework for combinatorial optimization.