Train Scheduling with Deep Q-Network: A Feasibility Test
We consider a train scheduling problem in which both local and express trains are to be scheduled. In this type of train scheduling problem, the key decision is determining the overtaking stations at which express trains overtake their preceding local trains. This problem has been successfully model...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-11-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/10/23/8367 |
| _version_ | 1797546852667621376 |
|---|---|
| author | Intaek Gong Sukmun Oh Yunhong Min |
| author_facet | Intaek Gong Sukmun Oh Yunhong Min |
| author_sort | Intaek Gong |
| collection | DOAJ |
| description | We consider a train scheduling problem in which both local and express trains are to be scheduled. In this type of train scheduling problem, the key decision is determining the overtaking stations at which express trains overtake their preceding local trains. This problem has been successfully modeled via mixed integer programming (MIP) models. One of the obvious limitation of MIP-based approaches is the lack of freedom to the choices objective and constraint functions. In this paper, as an alternative, we propose an approach based on reinforcement learning. We first decompose the problem into subproblems in which a single express train and its preceding local trains are considered. We, then, formulate the subproblem as a Markov decision process (MDP). Instead of solving each instance of MDP, we train a deep neural network, called deep <i>Q</i>-network (DQN), which approximates <i>Q</i>-value function of any instances of MDP. The learned DQN can be used to make decision by choosing the action which corresponds to the maximum <i>Q</i>-value. The advantage of the proposed method is the ability to incorporate any complex objective and/or constraint functions. We demonstrate the performance of the proposed method by numerical experiments. |
| first_indexed | 2024-03-10T14:35:06Z |
| format | Article |
| id | doaj.art-871df17b6bca4126a56214942d6b44e0 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T14:35:06Z |
| publishDate | 2020-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-871df17b6bca4126a56214942d6b44e02023-11-20T22:15:12ZengMDPI AGApplied Sciences2076-34172020-11-011023836710.3390/app10238367Train Scheduling with Deep Q-Network: A Feasibility TestIntaek Gong0Sukmun Oh1Yunhong Min2Graduate School of Logistics, Incheon National University, Incheon 22012, KoreaTransport System Research Team, Korea Railroad Research Institute, Gyeonggi 16105, KoreaGraduate School of Logistics, Incheon National University, Incheon 22012, KoreaWe consider a train scheduling problem in which both local and express trains are to be scheduled. In this type of train scheduling problem, the key decision is determining the overtaking stations at which express trains overtake their preceding local trains. This problem has been successfully modeled via mixed integer programming (MIP) models. One of the obvious limitation of MIP-based approaches is the lack of freedom to the choices objective and constraint functions. In this paper, as an alternative, we propose an approach based on reinforcement learning. We first decompose the problem into subproblems in which a single express train and its preceding local trains are considered. We, then, formulate the subproblem as a Markov decision process (MDP). Instead of solving each instance of MDP, we train a deep neural network, called deep <i>Q</i>-network (DQN), which approximates <i>Q</i>-value function of any instances of MDP. The learned DQN can be used to make decision by choosing the action which corresponds to the maximum <i>Q</i>-value. The advantage of the proposed method is the ability to incorporate any complex objective and/or constraint functions. We demonstrate the performance of the proposed method by numerical experiments.https://www.mdpi.com/2076-3417/10/23/8367train schedulingreinforcement learningdeep Q-network |
| spellingShingle | Intaek Gong Sukmun Oh Yunhong Min Train Scheduling with Deep Q-Network: A Feasibility Test Applied Sciences train scheduling reinforcement learning deep Q-network |
| title | Train Scheduling with Deep Q-Network: A Feasibility Test |
| title_full | Train Scheduling with Deep Q-Network: A Feasibility Test |
| title_fullStr | Train Scheduling with Deep Q-Network: A Feasibility Test |
| title_full_unstemmed | Train Scheduling with Deep Q-Network: A Feasibility Test |
| title_short | Train Scheduling with Deep Q-Network: A Feasibility Test |
| title_sort | train scheduling with deep q network a feasibility test |
| topic | train scheduling reinforcement learning deep Q-network |
| url | https://www.mdpi.com/2076-3417/10/23/8367 |
| work_keys_str_mv | AT intaekgong trainschedulingwithdeepqnetworkafeasibilitytest AT sukmunoh trainschedulingwithdeepqnetworkafeasibilitytest AT yunhongmin trainschedulingwithdeepqnetworkafeasibilitytest |