• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.2
• /
• pp.81-86
• /
• 2022

Given that time series are used in various fields, such as finance, IoT, and manufacturing, data analytical methods for accurate time-series forecasting can serve to increase operational efficiency. Among time-series analysis methods, multi-horizon forecasting provides a better understanding of data because it can extract meaningful statistics and other characteristics of the entire time-series. Furthermore, time-series data with exogenous information can be accurately predicted by using multi-horizon forecasting methods. However, traditional deep learning-based models for time-series do not account for the heterogeneity of inputs. We proposed an improved time-series predicting method, called the temporal fusion transformer method, which combines multi-horizon forecasting with interpretable insights into temporal dynamics. Various real-world data such as stock prices, fine dust concentrates and electricity consumption were considered in experiments. Experimental results showed that our temporal fusion transformer method has better time-series forecasting performance than existing models.

• Journal of the Korea Safety Management & Science
• /
• v.8 no.5
• /
• pp.139-150
• /
• 2006

The aim of this study is to establish an efficient distribution planning for a capacitated multi-stage supply chain. We assume that the demand information during planning horizon is given a deterministic form using a certain forecasting method. Under such a condition, we present a cost effective heuristic method for minimizing chain-wide supply chain inventory cost that is the sum of holding and backorder costs by using look-ahead technique. We cope with the capacity restriction constraints through look-ahead technique that considers not only the current demand information but also future demand information. To evaluate performance of the proposed heuristic method, we compared it with the extant research that utilizes echelon stock concept, under various supply chain settings.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.6
• /
• pp.2227-2235
• /
• 2018

Under the current policies and compensation rules in South Korea, Photovoltaic (PV) generation supplier can maximize the profit by combining PV generation with Energy Storage System (ESS). However, the existing operational strategy of ESS is not able to maximize the profit due to the limitation of ESS capacity. In this paper, new ESS scheduling algorithm is introduced by utilizing the System Marginal Price (SMP) and PV generation forecasting to maximize the profits of PV generation supplier. The proposed algorithm determines the charging time of ESS by ranking the charging schedule from low to high SMP when PV generation is more than enough to charge ESS. The discharging time of ESS is determined by ranking the discharging schedule from high to low SMP when ESS energy is not enough to maintain the discharging. To compensate forecasting error, the algorithm is updated every hour to apply the up-to-date information. The simulation is performed to verify the effectiveness of the proposed algorithm by using actual PV generation and ESS information.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.479-482
• /
• 2021

시계열 형태의 데이터는 다양한 분야에서 수집되고 응용되기 때문에 정확한 시계열 예측은 많은 분야에서 운영 효율성을 높일 수 있는 중요한 분석 방법으로 고려된다. 그중 다층 수평 예측은 사용자에게 전반적인 시계열 데이터 경향성을 제공할 수 있다. 하지만 다양한 정보를 포함하는 시계열 데이터는 데이터에 내재한 이질성(heterogeneity)까지 포괄적으로 고려한 방법을 통해서만 정확한 예측을 할 수 있다. 하지만 지금까지 많은 시계열 분석 모델들이 데이터의 이질성을 반영하지 못했다. 이러한 한계를 보완하고자 우리는 Temporal Fusion Transformer 모델을 사용하여 실생활과 밀접한 관련이 있는 데이터에 적용하여 이질성을 고려한 향상된 예측을 수행하였다. 실제, 주식 데이터와 미세 먼지 데이터와 같은 실생활 시계열 데이터에 적용하였고 실험 결과 기존 모델보다 Mean Squared Error(MSE)가 0.3487 낮은 것을 확인하였다.

• Journal of the Korea Society for Simulation
• /
• v.15 no.3
• /
• pp.39-47
• /
• 2006

In most of supply chain planning practices, the estimated demands, which are forecasted for each individual period in a forecasting window, are regarded as deterministic. But, in reality, the forecasted demands for the periods of a given horizon are stochastically distributed. Instead of using a safety stock, this study considers a direct control of service level by choosing the demand used in planning from the distributed forecasted demand values for the corresponding period. Using the demand quantile and echelon stock concept, we propose a simple but efficient heuristic algorithm for multi-echelon serial systems under service level constraints. Through a comprehensive simulation study, the proposed algorithm was shown to be very accurate compared with the optimal solutions.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.38 no.1
• /
• pp.73-80
• /
• 2018

In motorway traffic flow control, the conventional way based on real-time response has been changed into advanced way based on proactive response. Future traffic conditions over multiple time intervals are crucial input data for advanced motorway traffic flow control. It is necessary to overcome the uncertainty of the future state in order for forecasting multiple-period traffic volumes, as the number of uncertainty concurrently increase when the forecasting horizon expands. In this vein, multi-interval forecasting of traffic volumes requires a viable approach to conquer future uncertainties successfully. In this paper, a forecasting model is proposed which effectively addresses the uncertainties of future state based on the behaviors of temporal evolution of traffic volume states that intrinsically exits in the big past data. The model selects the past states from the big past data based on the state evolution of current traffic volumes, and then the selected past states are employed for estimating future states. The model was also designed to be suitable for data management systems in practice. Test results demonstrated that the model can effectively overcome the uncertainties over multiple time periods and can generate very reliable predictions in term of prediction accuracy. Hence, it is indicated that the model can be mounted and utilized on advanced data management systems.