ABSTRACT
Optimization of equipment assignments for oil spill responses entails several real-world constraints. We proposes a surrogate model which utilizes a deep neural network for the optimization of oil-removal equipment assignments. The surrogate model was constructed by applying machine-learning to 20,000 assignment plan data, all of which satisfy various constraining conditions based on deep neural networks. Compared to the existing optimization model, the constructed model showed a 61% increase in efficiency.
- Hye-Jin Kim, Junghwan Lee, Jong-Hwui Yun, and Yong-Hyuk Kim. 2019. Optimal equipment assignment for oil spill response using a genetic algorithm. In Proceedings of the Genetic and Evolutionary Computation Conference Companion. ACM, Prague Czech Republic, 375--376.Google ScholarDigital Library
- Jong-Hwui Yun, Dong-O Jo, Seunggi Guk, Yeongro Choi, Wondon Kim, Gyeong-Woo Jo, Dong-Hyeon Choi, Sang-Goo Kim, Jung-Hwan Moon, Ha-Yong Jang, Yeong-Nam Park, Eunmi Guk, and Eunbi Park. 2009. A Study on Practical Strategies for Estimating the National Control Ability of Oil Spill Control. Korea Maritime and Ocean University Technical Report. Korea Coast Guard.Google Scholar
Index Terms
- A surrogate model using deep neural networks for optimal oil skimmer assignment
Recommendations
Optimal equipment assignment for oil spill response using a genetic algorithm
GECCO '19: Proceedings of the Genetic and Evolutionary Computation Conference CompanionWe propose a genetic algorithm (GA)-based optimal equipment assignment method for oil spill response. We devised a repair operation suitable for constrained equipment assignment. In addition, the assignment strategies were evaluated by simulation to ...
Extending coupled hydrological-hydraulic model chains with a surrogate model for the estimation of flood losses
AbstractIn comparison to a local-scale flood risk analysis, modeling flood losses and risks at the river basin scale is challenging. Particularly in mountainous watersheds, extreme precipitation can be distributed spatially and temporally with ...
Graphical abstractDisplay Omitted
Highlights- A novel approach for coupling flood loss models with hydrological-hydraulic models.
Recursive surrogate model based on generalized regression neural network
AbstractSurrogate models have attracted considerable interest as approximation tools that can save considerable computational resources in various applications. In this study, a recursive surrogate model was developed based on a generalized ...
Highlights- A novel recursive correction method is proposed to continuously improve the prediction accuracy of GRNN model.
Comments