Deep reinforcement learning for selecting demand forecast models to empower Industry 3.5 and an empirical study for a semiconductor component distributor

2020 ◽  
Vol 58 (9) ◽  
pp. 2784-2804 ◽  
Author(s):  
Chen-Fu Chien ◽  
Yun-Siang Lin ◽  
Sheng-Kai Lin
Keyword(s):  
Reinforcement Learning ◽  
Empirical Study ◽  
Demand Forecast ◽  
Forecast Models

Multitask Learning and Reinforcement Learning for Personalized Dialog Generation: An Empirical Study

2021 ◽  
Vol 32 (1) ◽  
pp. 49-62
Author(s):  
Min Yang ◽  
Weiyi Huang ◽  
Wenting Tu ◽  
Qiang Qu ◽  
Ying Shen ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Empirical Study ◽  
Multitask Learning

scholarly journals Explaining Reinforcement Learning to Mere Mortals: An Empirical Study

2019 ◽  
Author(s):  
Andrew Anderson ◽  
Jonathan Dodge ◽  
Amrita Sadarangani ◽  
Zoe Juozapaitis ◽  
Evan Newman ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Empirical Study ◽  
Mental Models ◽  
Mental Model ◽  
User Study ◽  
Focus Of Attention ◽  
Saliency Maps ◽  
Treatment Experiment ◽  
Rts Game ◽  
The Impact

We present a user study to investigate the impact of explanations on non-experts? understanding of reinforcement learning (RL) agents. We investigate both a common RL visualization, saliency maps (the focus of attention), and a more recent explanation type, reward-decomposition bars (predictions of future types of rewards). We designed a 124 participant, four-treatment experiment to compare participants? mental models of an RL agent in a simple Real-Time Strategy (RTS) game. Our results show that the combination of both saliency and reward bars were needed to achieve a statistically significant improvement in mental model score over the control. In addition, our qualitative analysis of the data reveals a number of effects for further study.

Energy Consumption Forecasting in Hong Kong Using ARIMA and Artificial Neural Networks Models

2014 ◽  
Vol 672-674 ◽  
pp. 2085-2097 ◽  
Author(s):  
Sue Ling Lai ◽  
Ming Liu ◽  
Kuo Cheng Kuo ◽  
Ray Chang
Keyword(s):  
Hong Kong ◽  
Energy Consumption ◽  
Energy Demand ◽  
Moving Average ◽  
Critical Role ◽  
Forecast Accuracy ◽  
Demand Forecast ◽  
Forecast Models ◽  
Artificial Neural ◽  
Artificial Neural Network Ann

There have been considerable efforts contributed to the development of effective energy demand forecast models due to its critical role for economic development and environmental protection. This study focused on the adoption of artificial neural network (ANN) and autoregressive integrated moving average (ARIMA) models for energy consumption forecasting in Hong Kong over the period of 1975-2010. Four predictors were considered, including population, GDP, exports, and total visitor arrivals. The results show most ANN models demonstrate acceptable forecast accuracy when single predictor is considered. The best single input model is the case with GDP as predictor. Population and exports are the next proper single inputs. The model with total visitor arrivals as sole predictor does not perform satisfactorily. This indicates that tourism development demonstrates a different pattern from that of energy consumption. In addition, the forecast accuracy of ANN does not improve considerably as the number of predictors increase. Findings imply that with the ANN approach, choosing appropriate predictors is more important than increasing the number of predictors. On the other hand, ARIMA generates forecasts as accurate as some good cases by ANN. Results suggest that ARIMA is not only a parsimonious but effective approach for energy consumption forecasting in Hong Kong.

scholarly journals The application of demand forecast models: case study on bearing factory

2018 ◽  
Vol 4 (15) ◽  
Author(s):  
Bruno Santos Correa ◽  
Rosivan Cunha da Silva ◽  
Maílson Batista de Vilhena ◽  
Ana Paula de Souza e Silva
Keyword(s):  
Demand Forecast ◽  
Bearing Factory ◽  
Forecast Models

scholarly journals Energy Demand Forecast Models for Commercil Buildings in South Korea

Energies ◽  
2019 ◽  
Vol 12 (12) ◽  
pp. 2313 ◽  
Author(s):  
Sungkyun Ha ◽  
Sungho Tae ◽  
Rakhyun Kim
Keyword(s):  
South Korea ◽  
Energy Demand ◽  
Energy Model ◽  
Commercial Buildings ◽  
Demand Forecast ◽  
Annual Average ◽  
Average Growth ◽  
Energy Demands ◽  
City Gas ◽  
Forecast Models

With the Paris Agreement entering into full force, South Korea must submit its target greenhouse gas emissions for commercial buildings by 2030 to the United Nations Framework Convention on Climate Change. To determine this target, the annual energy demands must be forecasted through appropriate models; the development of these models is the focus of our study. We developed a system to calculate energy demand forecasts by searching for suitable methods. We built distinct energy forecast models for petroleum, city gas, electricity, heat, and renewable energies. The results show that the most appropriate variable for the petroleum energy model is energy trend. Moreover, the annual increase rate of petroleum energy demand from 2019 to 2030 was forecasted to be −1.7%. The appropriate variable for city gas energy model was the floor area of commercial buildings, which was forecasted to increase at an annual average growth rate of 0.4% from 2019 to 2030. According to the forecast results of energy demand from 2019 to 2030, the annual average growth rates of electricity, heat, and renewable energy demands were 2.1%, −0.2%, and 1.3%, respectively.

Sign in / Sign up

Export Citation Format

Share Document