Generating Diverse Optimal Road Management Plans in Post-Disaster by Applying Envelope Multi-Objective Deep Reinforcement Learning
-
Published:2023-12-01
Issue:8
Volume:18
Page:884-894
-
ISSN:1883-8030
-
Container-title:Journal of Disaster Research
-
language:en
-
Short-container-title:J. Disaster Res.
Author:
Joo Soo-Hyun1ORCID, Ogawa Yoshiki2ORCID, Sekimoto Yoshihide2ORCID
Affiliation:
1. International Development & Infrastructure Network Lab, Hongik University, Z1-202, 94 Wausan-ro, Mapo-gu, Seoul 04066, Korea 2. Center for Spatial Information Science, The University of Tokyo, Tokyo, Japan
Abstract
The authors used a data-driven reinforcement learning model for the post-disaster rapid recovery of human mobility, considering human-mobility recovery rate, road connectivity, and travel cost as the recovery components, to generate the reward framework. Each component has relative importance with respect to the others. However, if the preference is different from the original one, the optimal policy may not always be identified. This limitation must be addressed to enhance the robustness and generalizability of the proposed deep Q-network model. Therefore, a set of optimal policies were identified over a predetermined preference space, and the underlying importance was evaluated by applying envelope multi-objective reinforcement learning. The agent used in this study could distinguish the importance of each damaged road based on a given relative preference and derive a road-recovery policy suitable for each criterion. Furthermore, the authors provided the guidelines for constructing the optimal road-management plan. Based on the generalized policy network, the government can access diverse restoration strategies and select the most appropriate one depending on the disaster situation.
Publisher
Fuji Technology Press Ltd.
Subject
Engineering (miscellaneous),Safety, Risk, Reliability and Quality
Reference39 articles.
1. Weathernews, Inc., “Analysis of 20,000 flood damage reports in Western Japan Flooding,” July 10, 2018 (in Japanese). https://jp.weathernews.com/news/23807/ [Accessed November 1, 2022] 2. S. Joo, Y. Ogawa, and Y. Sekimoto, “Road-reconstruction after multi-locational flooding in multi-agent deep RL with the consideration of human mobility – Case study: Western Japan flooding in 2018,” Int. J. of Disaster Risk Reduction, Vol.70, 102780, 2022. https://doi.org/10.1016/j.ijdrr.2021.102780 3. Asahi Weekly, “582 sections of roads nationwide are closed and traffic jams occur,” July 15, 2018. https://www.asahi.com/articles/ASL7H35D0L7HPTIL00H.html [Accessed October 10, 2022] 4. R. Wako, Y. Sekimoto, H. Kanasugi, and R. Shibasaki, “Analysis of people’s route and destination choice in evacuation using GPS log data,” J. of Japan Society of Civil Engineering, Ser. D3 (Infrastructure Planning and Management), Vol.70, No.5, pp. I_681-I_688, 2014 (in Japanese). https://doi.org/10.2208/jscejipm.70.I_681 5. X. Lu, L. Bengtsson, and P. Holme, “Predictability of population displacement after the 2010 Haiti earthquake,” Proc. of the National Academy of Sciences, Vol.109, No.29, pp. 11576-11581, 2012. https://doi.org/10.1073/pnas.1203882109
|
|