| 1. |
- May, Ross
(författare)
-
The reinforcement learning method : A feasible and sustainable control strategy for efficient occupant-centred building operation in smart cities
- 2019
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Over half of the world’s population lives in urban areas, a trend which is expected to only grow as we move further into the future. With this increasing trend in urbanisation, challenges are presented in the form of the management of urban infrastructure systems. As an essential infrastructure of any city, the energy system presents itself as one of the biggest challenges. As cities expand in population and economically, global energy consumption increases and as a result so do greenhouse gas (GHG) emissions. To achieve the 2030 Agenda’s sustainable development goal on energy (SDG 7), renewable energy and energy efficiency have been shown as key strategies for attaining SDG 7. As the largest contributor to climate change, the building sector is responsible for more than half of the global final energy consumption and GHG emissions. As people spend most of their time indoors, the demand for energy is made worse as a result of maintaining the comfort level of the indoor environment. However, the emergence of the smart city and the internet of things (IoT) offers the opportunity for the smart management of buildings. Focusing on the latter strategy towards attaining SDG 7, intelligent building control offers significant potential for saving energy while respecting occupant comfort (OC). Most intelligent control strategies, however, rely on complex mathematical models which require a great deal of expertise to construct thereby costing in time and money. Furthermore, if these are inaccurate then energy is wasted and the comfort of occupants is decreased. Moreover, any change in the physical environment such as retrofits result in obsolete models which must be re-identified to match the new state of the environment. This model-based approach seems unsustainable and so a new model-free alternative is proposed. One such alternative is the reinforcement learning (RL) method. This method provides a beautiful solution to accomplishing the tradeoff between energy efficiency and OC within the smart city and more importantly to achieving SDG 7. To address the feasibility of RL as a sustainable control strategy for efficient occupant-centred building operation, a comprehensive review of RL for controlling OC in buildings as well as a case study implementing RL for improving OC via a window system are presented. The outcomes of each seem to suggest RL as a feasible solution, however, more work is required in the form of addressing current open issues such as cooperative multi-agent RL (MARL) needed for multi-occupant/multi-zonal buildings.
|
|
| 2. |
- Carling, Kenneth, et al.
(författare)
-
GRASP and statistical bounds for heuristic solutions to combinatorial problems
- 2016
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- The quality of a heuristic solution to a NP-hard combinatorial problem is hard to assess. A few studies have advocated and tested statistical bounds as a method for assessment. These studies indicate that statistical bounds are superior to the more widely known and used deterministic bounds. However, the previous studies have been limited to a few metaheuristics and combinatorial problems and, hence, the general performance of statistical bounds in combinatorial optimization remains an open question. This work complements the existing literature on statistical bounds by testing them on the metaheuristic Greedy Randomized Adaptive Search Procedures (GRASP) and four combinatorial problems. Our findings confirm previous results that statistical bounds are reliable for the p-median problem, while we note that they also seem reliable for the set covering problem. For the quadratic assignment problem, the statistical bounds has previously been found reliable when obtained from the Genetic algorithm whereas in this work they found less reliable. Finally, we provide statistical bounds to four 2-path network design problem instances for which the optimum is currently unknown.
|
|
| 3. |
- Carling, Kenneth, et al.
(författare)
-
Measuring CO2 emissions induced by online and brick-and-mortar retailing
- 2014
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- We develop a method for empirically measuring the difference in carbon footprint between traditional and online retailing (“e-tailing”) from entry point to a geographical area to consumer residence. The method only requires data on the locations of brick-and-mortar stores, online delivery points, and residences of the region’s population, and on the goods transportation networks in the studied region. Such data are readily available in most countries, so the method is not country or region specific. The method has been evaluated using data from the Dalecarlia region in Sweden, and is shown to be robust to all assumptions made. In our empirical example, the results indicate that the average distance from consumer residence to a brick-and-mortar retailer is 48.54 km in the studied region, while the average distance to an online delivery point is 6.7 km. The results also indicate that e-tailing increases the average distance traveled from the regional entry point to the delivery point from 47.15 km for a brick-and-mortar store to 122.75 km for the online delivery points. However, as professional carriers transport the products in bulk to stores or online delivery points, which is more efficient than consumers’ transporting the products to their residences, the results indicate that consumers switching from traditional to e-tailing on average reduce their CO2 footprints by 84% when buying standard consumer electronics products.
|
|
| 4. |
- Carlucci, S., et al.
(författare)
-
Modeling occupant behavior in buildings
- 2020
-
Ingår i: Building and Environment. - : Elsevier BV. - 0360-1323 .- 1873-684X. ; 174
-
Tidskriftsartikel (refereegranskat)
|
|
| 5. |
|
|
| 6. |
|
|
| 7. |
- Han, Mengjie, 1985-, et al.
(författare)
-
A novel reinforcement learning method for improving occupant comfort via window opening and closing
- 2020
-
Ingår i: Sustainable cities and society. - : Elsevier BV. - 2210-6707. ; 61
-
Tidskriftsartikel (refereegranskat)abstract
- An occupant's window opening and closing behaviour can significantly influence the level of comfort in the indoor environment. Such behaviour is, however, complex to predict and control conventionally. This paper, therefore, proposes a novel reinforcement learning (RL) method for the advanced control of window opening and closing. The RL control aims at optimising the time point for window opening/closing through observing and learning from the environment. The theory of model-free RL control is developed with the objective of improving occupant comfort, which is applied to historical field measurement data taken from an office building in Beijing. Preliminary testing of RL control is conducted by evaluating the control method’s actions. The results show that the RL control strategy improves thermal and indoor air quality by more than 90 % when compared with the actual historically observed occupant data. This methodology establishes a prototype for optimally controlling window opening and closing behaviour. It can be further extended by including more environmental parameters and more objectives such as energy consumption. The model-free characteristic of RL avoids the disadvantage of implementing inaccurate or complex models for the environment, thereby enabling a great potential in the application of intelligent control for buildings.
|
|
| 8. |
|
|
| 9. |
- Han, Mengjie, 1985-, et al.
(författare)
-
An Approach to Data Acquisition for Urban Building Energy Modeling Using a Gaussian Mixture Model and Expectation-Maximization Algorithm
- 2021
-
Ingår i: Buildings. - : MDPI. - 2075-5309. ; 11:1
-
Tidskriftsartikel (refereegranskat)abstract
- In recent years, a building’s energy performance is becoming uncertain because of factors such as climate change, the Covid-19 pandemic, stochastic occupant behavior and inefficient building control systems. Sufficient measurement data is essential to predict and manage a building’s performance levels. Assessing energy performance of buildings at an urban scale requires even larger data samples in order to perform an accurate analysis at an aggregated level. However, data are not only expensive, but it can also be a real challenge for communities to acquire large amounts of real energy data. This is despite the fact that inadequate knowledge of a full population will lead to biased learning and the failure to establish a data pipeline. Thus, this paper proposes a Gaussian mixture model (GMM) with an Expectation-Maximization (EM) algorithm that will produce synthetic building energy data. This method is tested on real datasets. The results show that the parameter estimates from the model are stable and close to the true values. The bivariate model gives better performance in classification accuracy. Synthetic data points generated by the models show a consistent representation of the real data. The approach developed here can be useful for building simulations and optimizations with spatio-temporal mapping.
|
|
| 10. |
- Han, Mengjie, 1985-, et al.
(författare)
-
Generating hourly electricity demand data for large-scale single-family buildings by a decomposition-recombination method
- 2022
-
Ingår i: Energy and Built Environment. - : Elsevier BV. - 2666-1233.
-
Tidskriftsartikel (refereegranskat)abstract
- Household electricity demand has substantial impacts on local grid operation, energy storage and the energy performance of buildings. Hourly demand data at district or urban level helps stakeholders understand the demand patterns from a granular time scale and provides robust evidence in energy management. However, such type of data is often expensive and time-consuming to collect, process and integrate. Decisions built upon smart meter data have to deal with challenges of privacy and security in the whole process. Incomplete data due to confidentiality concerns or system failure can further increase the difficulty of modeling and optimization. In addition, methods using historical data to make predictions can largely vary depending on data quality, local building environment, and dynamic factors. Considering these challenges, this paper proposes a statistical method to generate hourly electricity demand data for large-scale single-family buildings by decomposing time series data and recombining them into synthetics. The proposed method used public data to capture seasonality and the distribution of residuals that fulfill statistical characteristics. A reference building was used to provide empirical parameter settings and validations for the studied buildings. An illustrative case in a city of Sweden using only annual total demand was presented for deploying the proposed method. The results showed that the proposed method can mimic reality well and represent a high level of similarity to the real data. The average monthly error for the best month reached 15.9% and the best one was below 10% among 11 tested months. Less than 0.6% improper synthetic values were found in the studied region.
|
|
| 11. |
- Han, Mengjie, 1985-, et al.
(författare)
-
GRASP and Statistical Bounds for Heuristic Solutions to Combinatorial Problems
- 2019
-
Ingår i: International Journal of Management and Applied Science. - : Institute of Research and Journals (IRAJ). - 2394-7926. ; 5:8, s. 113-119
-
Tidskriftsartikel (refereegranskat)abstract
- The quality of a heuristic solution to a NP-hard combinatorial problem is hard to assess. A few studies have advocated and tested statistical bounds as a method for assessment. These studies indicate that statistical bounds are superior to the more widely known and used deterministic bounds. However, the previous studies have been limited to a few heuristics and combinatorial problems and, hence, the general performance of statistical bounds in combinatorial optimization remains an open question. This work complements the existing literature on statistical bounds by testing them on the metaheuristic Greedy Randomized Adaptive Search Procedures (GRASP) and four combinatorial problems. Our findings confirm previous results that statistical bounds are reliable for the p-median problem, while we note that they also seem reliable for the set covering problem. For the quadratic assignment problem, the statistical bounds have previously been found reliable when obtained from the Genetic algorithm whereas in this work they have been found less reliable. Finally, we provide statistical bounds to four 2-path network design problem instances for which the optimum is currently unknown.
|
|
| 12. |
- Han, Mengjie, 1985-
(författare)
-
Heuristic optimization of the p-median problem and population re-distribution
- 2013
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis contributes to the heuristic optimization of the p-median problem and Swedish population redistribution. The p-median model is the most representative model in the location analysis. When facilities are located to a population geographically distributed in Q demand points, the p-median model systematically considers all the demand points such that each demand point will have an effect on the decision of the location. However, a series of questions arise. How do we measure the distances? Does the number of facilities to be located have a strong impact on the result? What scale of the network is suitable? How good is our solution? We have scrutinized a lot of issues like those. The reason why we are interested in those questions is that there are a lot of uncertainties in the solutions. We cannot guarantee our solution is good enough for making decisions. The technique of heuristic optimization is formulated in the thesis. Swedish population redistribution is examined by a spatio-temporal covariance model. A descriptive analysis is not always enough to describe the moving effects from the neighbouring population. A correlation or a covariance analysis is more explicit to show the tendencies. Similarly, the optimization technique of the parameter estimation is required and is executed in the frame of statistical modeling.
|
|
| 13. |
- Han, Mengjie, 1985-, et al.
(författare)
-
Intra-urban location of stores and labour turnover in retail
- 2019
-
Ingår i: International Review of Retail Distribution & Consumer Research. - : Informa UK Limited. - 0959-3969 .- 1466-4402. ; 29:4, s. 359-375
-
Tidskriftsartikel (refereegranskat)abstract
- The aim of this paper is to analyse labour turnover in retail firms with stores in different city locations. This case study of a Swedish mid-sized city uses comprehensive longitudinal register data on individuals. In a first step, an unconditional descriptive analysis shows that labour turnover in retail is higher in out-of-town locations, compared to more central locations in the city. In a second step, a generalized linear model (GLM) analysis is conducted where labour turnover in downtown and out-of-town locations are compared. Firm internal and industry factors, as well as employee characteristics, and location-specific factors are controlled for. The results indicate that commuting costs and intra-urban location have no statistically significant effect on labour turnover in retail firms. Instead, firm internal factors, such as human resource management, has a major influence on labour turnover rates. The findings indicate that in particular firms with multiple locations may need to pay extra attention to work conditions across stores in different places in a city, in order to avoid diverging levels of labour mobility. This paper complements previous survey-based studies on labour turnover by using a comprehensive micro-level dataset to analyse revealed rather than stated preferences concerning job-to-job mobility. An elaborated measure of labour turnover is used to analyse differences between shopping areas in different locations within the city. The particular research design used in this paper makes it possible to isolate the effect of intra-organizational conditions by analysing mobility within firms with workplaces in both downtown and out-of-town locations. This is the first comprehensive study of labour turnover and mobility with an intra-urban perspective in the retail sector.
|
|
| 14. |
- Han, Mengjie, 1985-, et al.
(författare)
-
Perspectives of Machine Learning and Natural Language Processing on Characterizing Positive Energy Districts
- 2024
-
Ingår i: Buildings. - : MDPI. - 2075-5309. ; 14:2
-
Tidskriftsartikel (refereegranskat)abstract
- The concept of a Positive Energy District (PED) has become a vital component of the efforts to accelerate the transition to zero carbon emissions and climate-neutral living environments. Research is shifting its focus from energy-efficient single buildings to districts, where the aim is to achieve a positive energy balance across a given time period. Various innovation projects, programs, and activities have produced abundant insights into how to implement and operate PEDs. However, there is still no agreed way of determining what constitutes a PED for the purpose of identifying and evaluating its various elements. This paper thus sets out to create a process for characterizing PEDs. First, nineteen different elements of a PED were identified. Then, two AI techniques, machine learning (ML) and natural language processing (NLP), were introduced and examined to determine their potential for modeling, extracting, and mapping the elements of a PED. Lastly, state-of-the-art research papers were reviewed to identify any contribution they can make to the determination of the effectiveness of the ML and NLP models. The results suggest that both ML and NLP possess significant potential for modeling most of the identified elements in various areas, such as optimization, control, design, and stakeholder mapping. This potential is realized through the utilization of vast amounts of data, enabling these models to generate accurate and useful insights for PED planning and implementation. Several practical strategies have been identified to enhance the characterization of PEDs. These include a clear definition and quantification of the elements, the utilization of urban-scale energy modeling techniques, and the development of user-friendly interfaces capable of presenting model insights in an accessible manner. Thus, developing a holistic approach that integrates existing and novel techniques for PED characterization is essential to achieve sustainable and resilient urban environments.
|
|
| 15. |
|
|
| 16. |
- Han, Mengjie, 1985-, et al.
(författare)
-
Review of natural language processing techniques for characterizing positive energy districts
- 2023
-
Ingår i: journal of Physics; Conference series. - : Institute of Physics Publishing (IOPP).
-
Konferensbidrag (refereegranskat)abstract
- The concept of Positive Energy Districts (PEDs) has emerged as a crucial aspect of endeavours aimed at accelerating the transition to zero carbon emissions and climate-neutral living spaces. The focus of research has shifted from energy-efficient individual buildings to entire districts, where the objective is to achieve a positive energy balance over a specific timeframe. The consensus on the conceptualization of a PED has been evolving and a standardized checklist for identifying and evaluating its constituent elements needs to be addressed. This study aims to develop a methodology for characterizing PEDs by leveraging natural language processing (NLP) techniques to model, extract, and map these elements. Furthermore, a review of state-of-the-art research papers is conducted to ascertain their contribution to assessing the effectiveness of NLP models. The findings indicate that NLP holds significant potential in modelling the majority of the identified elements across various domains. To establish a systematic framework for AI modelling, it is crucial to adopt approaches that integrate established and innovative techniques for PED characterization. Such an approach would enable a comprehensive and effective implementation of NLP within the context of PEDs, facilitating the creation of sustainable and resilient urban environments. © 2023 Institute of Physics Publishing. All rights reserved.
|
|
| 17. |
- Han, Mengjie, 1985-, et al.
(författare)
-
The reinforcement learning method for occupant behavior in building control : A review
- 2021
-
Ingår i: Energy and Built Environment. - : Elsevier BV. - 2666-1233. ; 2:2, s. 137-148
-
Tidskriftsartikel (refereegranskat)abstract
- Occupant behavior in buildings has been considered the major source of uncertainty for assessing energy consumption and building performance. Modeling frameworks are usually built to accomplish a certain task, but the stochasticity of the occupant makes it difficult to apply that experience to a similar but distinct environment. For complex and dynamic environments, the development of smart devices and computing power makes intelligent control methods for occupant behaviors more viable. It is expected that they will make a substantial contribution to reducing global energy consumption. Among these control techniques, the reinforcement learning (RL) method seems distinctive and applicable. The success of the reinforcement learning method in many artificial intelligence applications has given an explicit indication of how this method might be used to model and adjust occupant behavior in building control. Fruitful algorithms complement each other and guarantee the quality of the optimization. However, the examination of occupant behavior based on reinforcement learning methodologies is not well established. The way that occupant interacts with the RL agent is still unclear. This study briefly reviews the empirical applications using reinforcement learning, how they have contributed to shaping the modeling paradigms and how they might suggest a future research direction.
|
|
| 18. |
- Huang, Pei, et al.
(författare)
-
Characterization and optimization of energy sharing performances in energy-sharing communities in Sweden, Canada and Germany
- 2022
-
Ingår i: Applied Energy. - : Elsevier. - 0306-2619 .- 1872-9118. ; 326
-
Tidskriftsartikel (refereegranskat)abstract
- Peer-to-peer (P2P) renewable power sharing within a building community is a promising solution to enhance the community's self-sufficiency and relieve the grid stress posed by the increased deployment of distributed renewable power. Existing studies have pointed out that the energy sharing potentials of a building community are affected by various factors including location, community scale, renewable energy system (RES) capacity, energy system type, storage integration, etc. However, the impacts of these factors on the energy sharing potentials in a building community are not fully studied. Being unaware of those factors’ impacts could lead to reduced energy sharing potentials and thus limit the associated improvement in energy and economic performances. Thus, this study conducts a comprehensive analysis of various factors’ impacts on the energy sharing performances in building communities. Two performance indicators are first proposed to quantify the energy sharing performances: total amount of energy sharing and energy sharing ratio (ESR). Then, parametric studies are conducted based on real electricity demand data in three countries to reveal how these factors affect the proposed indictors and improvements in self-sufficiency, electricity costs, and energy exchanges with the power grid. Next, a genetic algorithm based design method is developed to optimize the influential parameters to maximize the energy sharing potentials in a community. The study results show that the main influential factors are RES capacity ratio, PV capacity ratio, and energy storage system capacity. A large energy storage capacity can enhance the ESR. To achieve the maximized ESR, the optimal RES capacity ratio should be around 0.4 ∼ 1.1. The maximum energy sharing ratio is usually smaller in high latitude districts such as Sweden. This study characterizes the energy sharing performances and provides a novel perspective to optimize the design of energy systems in energy sharing communities. It can pave the way for the large integration of distributed renewable power in the future. © 2022 The Author(s)
|
|
| 19. |
|
|
| 20. |
|
|
| 21. |
- Jin, Y., et al.
(författare)
-
District household electricity consumption pattern analysis based on auto-encoder algorithm
- 2019
-
Ingår i: IOP Conference Series: Materials Science and Engineering.
-
Konferensbidrag (refereegranskat)abstract
- The energy shortage is one key issue for sustainable development, a potential solution of which is the integration with the renewable energy resources. However, the temporal sequential characteristic of renewable resources is different from traditional power grid. For the entire power grid, it is essential to match the energy generation side with the energy consumption side, so the load characteristic at the energy use side is crucial for renewable power integration. Better understanding of energy consumption pattern in buildings contributes to matching different source of energy generation. Under the background of integration of traditional and renewable energy, this research focuses on analysis of different household electricity consumption patterns in an urban scale. The original data is from measurement of daily energy consumption with smart meter in households. To avoid the dimension explosion phenomenon, the auto-encoder algorithm is introduced during the clustering analysis of daily electricity use data, which plays the role of principal component analysis. The clustering based on auto-encoder gives a clear insight into the urban electricity use patterns in household. During the data analysis, several feature variables are proposed, which include peak value, valley value and average value. The distinction analysis is also conducted to evaluate the analysis performance. The study takes households in Nanjing city, China as a case study, to conduct the clustering analysis on electricity consumption of residential buildings. The analysis results can be further applied, such as during the capacity design of district energy storage.
|
|
| 22. |
|
|
| 23. |
- May, Ross
(författare)
-
On the Feasibility of Reinforcement Learning in Single- and Multi-Agent Systems : The Cases of Indoor Climate and Prosumer Electricity Trading Communities
- 2023
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Over half of the world’s population live in urban areas, a trend which is expected to only grow as we move further into the future. With this increasing trend in urbanisation, challenges are presented in the form of the management of urban infrastructure systems. As an essential infrastructure of any city, the energy system presents itself as one of the biggest challenges. Indeed, as cities expand in population and economically, global energy consumption increases, and as a result, so do greenhouse gas (GHG) emissions. Key to realising the goals as laid out by the 2030 Agenda for Sustainable Development, is the energy transition - embodied in the goals pertaining to affordable and clean energy, sustainable cities and communities, and climate action. Renewable energy systems (RESs) and energy efficiency have been shown as key strategies towards achieving these goals. While the building sector is considered to be one of the biggest contributors to climate change, it is also seen as an area with many opportunities for realising the energy transition. Indeed, the emergence of the smart city and the internet of things (IoT), alongside Photovoltaic and battery technology, offers opportunities for both the smart management of buildings, as well as the opportunity to form self-sufficient peer-to-peer (P2P) electricity trading communities. Within this context, advanced building control offers significant potential for mitigating global warming, grid instability, soaring energy costs, and exposure to poor indoor building climates. Most advanced control strategies, however, rely on complex mathematical models, which require a great deal of expertise to construct, thereby costing in time and money, and are unlikely to be frequently updated - which can lead to sub-optimal or even wrong performance. Furthermore, arriving at solutions in economic settings as complex and dynamic as the P2P electricity markets referred to above, often leads to solutions that are computationally intractable. A model-based approach thus seems, as alluded to above, unsustainable, and I thus propose taking a model-free alternative instead. One such alternative is the reinforcement learning (RL) method. This method provides a beautiful solution that addresses many of the limitations seen in more classical approaches - those based on complex mathematical models - to single- and multi-agent systems. To address the feasibility of RL in the context of building systems, I have developed four papers. In studying the literature, while there is much review work in support of RL for controlling energy consumption, it was found that there were no such works analysing RL from a methodological perspective w.r.t. controlling the comfort level of building occupants. Thus, in Paper I, to fill in this gap in knowledge, a comprehensive review in this area was carried out. To follow up, in Paper II, a case study was conducted to further assess, among other things, the computational feasibility of RL for controlling occupant comfort in a single agent context. It was found that the RL method was able to improve thermal and indoor air quality by more than 90% when compared with historically observed occupant data. Broadening the scope of RL, Papers III and IV considered the feasibility of RL at the district scale by considering the efficient trade of renewable electricity in a peer-to-peer prosumer energy market. In particular, in Paper III, by extending an open source economic simulation framework, multi-agent reinforcement learning (MARL) was used to optimise a dynamic price policy for trading the locally produced electricity. Compared with a benchmark fixed price signal, the dynamic price mechanism arrived at by RL, increased community net profit by more than 28%, and median community self-sufficiency by more than 2%. Furthermore, emergent social-economic behaviours such as changes in supply w.r.t changes in price were identified. A limitation of Paper III, however, is that it was conducted in a single environment. To address this limitation and to assess the general validity of the proposed MARL-solution, in Paper IV a full factorial experiment based on the factors of climate - manifested in heterogeneous demand/supply profiles and associated battery parameters, community scale, and price mechanism, was conducted in order to ascertain the response of the community w.r.t net-loss (financial gain), self-sufficiency, and income equality from trading locally produced electricity. The central finding of Paper IV was that the community, w.r.t net-loss, performs significantly better under a learned dynamic price mechanism than under the benchmark fixed price mechanism, and furthermore, a community under such a dynamic price mechanism stands an odds of 2 to 1 in increased financial savings.
|
|
| 24. |
- May, Ross, et al.
(författare)
-
Reinforcement learning control for indoor comfort : A survey
- 2019
-
Ingår i: IOP Conference Series: Materials Science and Engineering.
-
Konferensbidrag (refereegranskat)abstract
- Building control systems are prone to fail in complex and dynamic environments. The reinforcement learning (RL) method is becoming more and more attractive in automatic control. The success of the reinforcement learning method in many artificial intelligence applications has resulted in an open question on how to implement the method in building control systems. This paper therefore conducts a comprehensive review of the RL methods applied in control systems for indoor comfort and environment. The empirical applications of RL-based control systems are then presented, depending on optimisation objectives and the measurement of energy use. This paper illustrates the class of algorithms and implementation details regarding how the value functions have been represented and how the policies are improved. This paper is expected to clarify the feasible theory and functions of RL for building control systems, which would promote their wider-spread application and thus contribute to the social economic benefits in the energy and built environments.
|
|
| 25. |
- Ngoc Phuong, Chau
(författare)
-
Machine Learning Approaches to Develop Weather Normalize Models for Urban Air Quality
- 2024
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- According to the World Health Organization, almost all human population (99%) lives in 117 countries with over 6000 cities, where air pollutant concentration exceeds recommended thresholds. The most common, so-called criteria, air pollutants that affect human lives, are particulate matter (PM) and gas-phase (SO2, CO, NO2, O3 and others). Therefore, many countries or regions worldwide have imposed regulations or interventions to reduce these effects. Whenever an intervention occurs, air quality changes due to changes in ambient factors, such as weather characteristics and human activities. One approach for assessing the effects of interventions or events on air quality is through the use of the Weather Normalized Model (WNM). However, current deterministic models struggle to accurately capture the complex, non-linear relationship between pollutant concentrations and their emission sources. Hence, the primary objective of this thesis is to examine the power of machine learning (ML) and deep learning (DL) techniques to develop and improve WNMs. Subsequently, these enhanced WNMs are employed to assess the impact of events on air quality. Furthermore, these ML/DL-based WNMs can serve as valuable tools for conducting exploratory data analysis (EDA) to uncover the correlations between independent variables (meteorological and temporal features) and air pollutant concentrations within the models. It has been discovered that DL techniques demonstrated their efficiency and high performance in different fields, such as natural language processing, image processing, biology, and environment. Therefore, several appropriate DL architectures (Long Short-Term Memory - LSTM, Recurrent Neural Network - RNN, Bidirectional Recurrent Neural Network - BIRNN, Convolutional Neural Network - CNN, and Gated Recurrent Unit - GRU) were tested to develop the WNMs presented in Paper I. When comparing these DL architectures and Gradient Boosting Machine (GBM), LSTM-based methods (LSTM, BiRNN) have obtained superior results in developing WNMs. The study also showed that our WNMs (DL-based) could capture the correlations between input variables (meteorological and temporal variables) and five criteria contaminants (SO2, CO, NO2, O3 and PM2.5). This is because the SHapley Additive exPlanations (SHAP) library allowed us to discover the significant factors in DL-based WNMs. Additionally, these WNMs were used to assess the air quality changes during COVID-19 lockdown periods in Ecuador. The existing normalized models operate based on the original units of pollutants and are designed for assessing pollutant concentrations under “average” or consistent weather conditions. Predicting pollution peaks presents an even greater challenge because they often lack discernible patterns. To address this, we enhanced the Weather Normalized Models (WNMs) to boost their performance specifically during daily concentration peak conditions. In the second paper, we accomplished this by developing supervised learning techniques, including Ensemble Deep Learning methods, to distinguish between daily peak and non-peak pollutant concentrations. This approach offers flexibility in categorizing pollutant concentrations as either daily concentration peaks or non-daily concentration peaks. However, it is worth noting that this method may introduce potential bias when selecting non-peak values. In the third paper, WNMs are directly applied to daily concentration peaks to predict and analyse the correlations between meteorological, temporal features and daily concentration peaks of air pollutants.
|
|
| 26. |
- Quintana, Samer, et al.
(författare)
-
A Top-Down Digital Mapping of Spatial-Temporal Energy Use for Municipality-Owned Buildings : A Case Study in Borlänge, Sweden
- 2021
-
Ingår i: Buildings. - : MDPI. - 2075-5309. ; 11:2
-
Tidskriftsartikel (refereegranskat)abstract
- Urban energy mapping plays a crucial role in benchmarking the energy performance of buildings for many stakeholders. This study examined a set of buildings in the city of Borlange, Sweden, owned by the municipality. The aim was to present a digital spatial map of both electricity use and district heating demand in the spatial-temporal dimension. A toolkit for top-down data processing and analysis was considered based on the energy performance database of municipality-owned buildings. The data were initially cleaned, transformed and geocoded using custom scripts and an application program interface (API) for OpenStreetMap and Google Maps. The dataset consisted of 228 and 105 geocoded addresses for, respectively, electricity and district heating monthly consumption for the year 2018. A number of extra parameters were manually incorporated to this data, i.e., the total floor area, the building year of construction and occupancy ratio. The electricity use and heating demand in the building samples were about 24.47 kWh/m(2) and 268.78 kWh/m(2), respectively, for which great potential for saving heating energy was observed. Compared to the electricity use, the district heating showed a more homogenous pattern following the changes of the seasons. The digital mapping revealed a spatial representation of identifiable hotspots for electricity uses in high-occupancy/density areas and for district heating needs in districts with buildings mostly constructed before 1980. These results provide a comprehensive means of understanding the existing energy distributions for stakeholders and energy advisors. They also facilitate strategy geared towards future energy planning in the city, such as energy benchmarking policies.
|
|
| 27. |
- Sadeghian, Paria
(författare)
-
A Multi-Dimensional Approach to Human Mobility and Transportation Mode Detection Using GPS Data
- 2024
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- GPS tracking data is an essential resource for analyzing human travel patterns and evaluating the effects on transportation systems. The primary challenge, however, is to accurately identify the modes of transportation within unlabeled GPS data. These approaches range from simple rule-based systems to advanced machine-learning techniques. This dissertation aims to bridge this gap by examining the critical features and techniques of these methods and proposing a novel approach for detecting transportation modes in GPS tracking data. To achieve this goal, a comprehensive understanding of individual journeys is crucial. Thus, this research adopts a microdata analytic approach, encompassing data collection, processing, analysis, and decision-making stages. Doing so contributes to advancing human mobility research and transportation mode detection. Paper I undertook a systematic review of transport mode detection methodologies to fill the research gap, emphasizing the predominance of supervised learning algorithms and highlighting the need for further research to address the limitations of small datasets. Paper II introduced a stepwise methodology, integrating unsupervised learning, GIS, and supervised algorithms to detect transport modes while minimizing reliance on labelled data. The Random Forest algorithm emerged as a precise but time-intensive solution. Paper III showcased a novel approach to transport mode detection using deep learning models, outperforming traditional machine learning methods. This paper signals the potential of deep learning in the field and demonstrates the importance of raw GPS data in enhancing accuracy. Paper V addressed the challenge of predicting human mobility patterns under the Hidden Markov Model (HMM) framework, highlighting the applicability of HMMs to understanding and predicting complex mobility behaviour. This paper emphasized the need for GPS tracking data in developing advanced mobility models. Paper IV ventured into hybrid methodology by combining K-means clustering with the ANP-PSO algorithm to enhance transportation mode classification. This pioneering approach improved classification accuracy while reducing dependence on labelled datasets. Collectively, these papers underscore the opportunities and limitations in human mobility research, offering insights into future directions for mitigating data quality issues and improving the accuracy of transportation mode detection. These innovative methodologies have practical implications for transportation planning, resource allocation, and intelligent transportation system development, ultimately shaping the future of transportation research and decision-making. Standardized data collection, processing, and labelling methods are crucial and need attention in future research. Future research can focus on developing such benchmarks and validation protocols to enhance the reliability and comparability of results.
|
|
| 28. |
- Sadeghian, Paria, et al.
(författare)
-
Testing feasibility of using a hidden Markov model on predicting human mobility based on GPS tracking data
- 2024
-
Ingår i: Transportmetrica B. - : Taylor & Francis. - 2168-0566. ; 12:1
-
Tidskriftsartikel (refereegranskat)abstract
- Human mobility behaviour is far from random and can be predictable. Predicting human mobility behaviour has the potential to improve location selection for facilities, transportation services, urban planning, and can be beneficial in providing more efficient sustainable urban development strategies. However, it is difficult to model urban mobility patterns since incentives for mobility is complex, and influenced by several factors, such as dynamic population, weather conditions. Thus, this paper proposes a prediction-oriented algorithm under the framework of a Hidden Markov Model to predict next-location and time-of-arrival of human mobility. A comprehensive evaluation of these two schemes for the representation of latent and observable variables is discussed. In conclusion, the paper provides a valuable contribution to the field of mobility behaviour prediction by proposing a novel algorithm. The evaluation shows that the proposed algorithm is stable and consistent in predicting the next location of users based on their past trajectories. © 2024 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
|
|
| 29. |
- Salin, Hannes, et al.
(författare)
-
Quality Metrics for Software Development Management and Decision Making : An Analysis of Attitudes and Decisions
- 2022
-
Ingår i: Product-Focused Software Process Improvement. 23rd International Conference, PROFES 2022, Jyväskylä, Finland, November 21–23, 2022, Proceedings. - Cham : Springer. - 9783031213885 - 9783031213878 ; , s. 525-530
-
Konferensbidrag (refereegranskat)abstract
- We combine current literature in software quality metrics with an attitude validation study with industry practitioners, to establish how quality metrics can be used for data-driven approaches. We also propose a simple metric nomenclature and map our findings into a decision making model for easy adoption and utilization of data-driven decision-making methods. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
|
|
| 30. |
- Shah, Juveria, et al.
(författare)
-
Analysis And Performance Mapping Of “Component To System” For A Parabolic Trough Collector Applied To Process Heating Applications
- 2022
-
Ingår i: ISEC 2022. ; , s. 487-488
-
Konferensbidrag (refereegranskat)abstract
- The slogan “Heat is half” is of importance to keep in mind that nearly 50 % of the final energy use is in the form of heat. The global efforts for future decarbonised heating systems are based on hydrogen and electrification of heating etc. Solar thermal technology is a key component of greener industrial heating solutions. Solar thermal technologies for process heating application has decade long history of implementation and are gaining significant interest from all around the world. The performance prediction of solar thermal technologies on the system level is more complicated compared to photovoltaic, due to the effect of performance on system boundary conditions such as variation in meteorological parameters, load demand, temperature levels, thermal storage type. The central focus of this paper is on the use of a parabolic trough collector (PTC) for process heating applications in the medium temperature range. The aim of this paper is to map the performance of PTC collector into an industrial system, and to analyse the decrease in collector thermal output from component level to system level. The simulations are implemented in TRNSYS and MATLAB. The results are visualized using QGIS tool to generate the heat map for performance parameters for a range of solar fractions.
|
|
| 31. |
- Wan, Benting, et al.
(författare)
-
An integrated group decision-making method for the evaluation of hypertension follow-up systems using interval-valued q-rung orthopair fuzzy sets
- 2023
-
Ingår i: Complex & Intelligent Systems. - : Springer Science and Business Media LLC. - 2199-4536 .- 2198-6053. ; 9:4, s. 4521-4554
-
Tidskriftsartikel (refereegranskat)abstract
- It is imperative to comprehensively evaluate the function, cost, performance and other indices when purchasing a hypertension follow-up (HFU) system for community hospitals. To select the best software product from multiple alternatives, in this paper, we develop a novel integrated group decision-making (GDM) method for the quality evaluation of the system under the interval-valued q-rung orthopair fuzzy sets (IVq-ROFSs). The design of our evaluation indices is based on the characteristics of the HFU system, which in turn represents the evaluation requirements of typical software applications and reflects the particularity of the system. A similarity is extended to measure the IVq-ROFNs, and a new score function is devised for distinguishing IVq-ROFNs to figure out the best IVq-ROFN. The weighted fairly aggregation (WFA) operator is then extended to the interval-valued q-rung orthopair WFA weighted average operator (IVq-ROFWFAWA) for aggregating information. The attribute weights are derived using the LINMAP model based on the similarity of IVq-ROFNs. We design a new expert weight deriving strategy, which makes each alternative have its own expert weight, and use the ARAS method to select the best alternative based on these weights. With these actions, a GDM algorithm that integrates the similarity, score function, IVq-ROFWFAWA operator, attribute weights, expert weights and ARAS is proposed. The applicability of the proposed method is demonstrated through a case study. Its effectiveness and feasibility are verified by comparing it to other state-of-the-art methods and operators.
|
|
| 32. |
- Wan, Benting, et al.
(författare)
-
Weighted average LINMAP group decision-making method based on q-rung orthopair triangular fuzzy numbers
- 2022
-
Ingår i: Granular Computing. - : Springer Nature. - 2364-4966 .- 2364-4974. ; 7:3, s. 489-503
-
Tidskriftsartikel (refereegranskat)abstract
- Considering the situation where decision values are q-rung orthopair triangular fuzzy number (q-ROTFN) and pair-wise comparisons of alternatives and evaluation matrices are given by decision-makers, a new group decision-making method is necessary to be studied for solving a group decision-making problem in the above situation. In this paper, we firstly proposed a q-rung orthopair triangular fuzzy weighted average (q-ROTFWA) operator based on the WA operator. In a second step, a linear programming technique for the multidimensional analysis of preferences (LINMAP) model based on q-ROTFN was formulated, which is used to obtain the weight of each attribute through partial preference information. A distance formula was introduced to get the ranking order of schemes and the best alternative. Finally, the weighted average LINMAP (WA-LINMAP) method was illustrated in a case study to verify its effectiveness. It is found in the experiment that the change of the q value does not affect the ranking of the schemes. The comparative analysis further confirms the effectiveness and feasibility of the proposed method.
|
|
| 33. |
- Wang, Zhenwu, et al.
(författare)
-
A multi-objective chicken swarm optimization algorithm based on dual external archive with various elites
- 2023
-
Ingår i: Applied Soft Computing. - : Elsevier BV. - 1568-4946 .- 1872-9681. ; 133
-
Tidskriftsartikel (refereegranskat)abstract
- Multi-objective optimization problems (MOPs) that widely exist in real world concern all optimal solutions compromised among multiple objectives. Chicken swarm optimization algorithm derived from emergent behaviors of organisms provides an effective way for handling MOPs. To speed up convergence and improve uniformity of Pareto-optimal solutions, a multi-objective chicken swarm optimization algorithm based on dual external archives and boundary learning strategy (MOCSO-DABL) is proposed in this paper. Dual external archives are employed to distinguish and choose two types of elite solutions, with the purpose of more effectively guiding individual evolution. A boundary learning strategy guides the chickens to learn from boundary individuals in the later stage of evolution. Moreover, fast non-dominated sorting is adopted to establish the hierarchical social structure of a chicken population, and learning strategies of roosters, hens and chicks are improved to meet the requirements of MOPs. Experimental results on 14 benchmark functions show that the proposed MOCSO-DABL outperforms other five state-of-the-art algorithms significantly.
|
|
| 34. |
- Wang, Zhenwu, et al.
(författare)
-
A novel bayesian network-based ensemble classifier chains for multi-label classification
- 2024
-
Ingår i: Complex & Intelligent Systems. - : Springer Berlin/Heidelberg. - 2199-4536 .- 2198-6053.
-
Tidskriftsartikel (refereegranskat)abstract
- In this paper, we address the challenges of random label ordering and limited interpretability associated with Ensemble Classifier Chains (ECC) by introducing a novel ECC method, ECC-MOO&BN, which integrates Bayesian Networks (BN) and Multi-Objective Optimization (MOO). This approach is designed to concurrently overcome these ECC limitations. The ECC-MOO&BN method focuses on extracting diverse and interpretable label orderings for the ECC classifier. We initiated this process by employing mutual information to investigate label relationships and establish the initial structures of the BN. Subsequently, an enhanced NSGA-II algorithm was applied to develop a series of Directed Acyclic Graphs (DAGs) that effectively balance the likelihood and complexity of the BN structure. The rationale behind using the MOO method lies in its ability to optimize both complexity and likelihood simultaneously, which not only diversifies DAG generation but also helps avoid overfitting during the production of label orderings. The DAGs, once sorted topologically, yielded a series of label orderings, which were then seamlessly integrated into the ECC framework for addressing multi-label classification (MLC) problems. Experimental results show that when benchmarked against eleven leading-edge MLC algorithms, our proposed method achieves the highest average ranking across seven evaluation criteria on nine out of thirteen MLC datasets. The results of the Friedman test and Nemenyi test also indicate that the performance of the proposed method has a significant advantage compared to other algorithms.
|
|
| 35. |
- Wang, Zhenwu, et al.
(författare)
-
A Three-Dimensional Visualization Framework forUnderground Geohazard Recognition on UrbanRoad-Facing GPR Data
- 2020
-
Ingår i: ISPRS International Journal of Geo-Information. - : MDPI AG. - 2220-9964. ; 9:11
-
Tidskriftsartikel (refereegranskat)abstract
- The identification of underground geohazards is always a difficult issue in the field of underground public safety. This study proposes an interactive visualization framework for underground geohazard recognition on urban roads, which constructs a whole recognition workflow by incorporating data collection, preprocessing, modeling, rendering and analyzing. In this framework, two proposed sampling point selection methods have been adopted to enhance the interpolated accuracy for the Kriging algorithm based on ground penetrating radar (GPR) technology. An improved Kriging algorithm was put forward, which applies a particle swarm optimization (PSO) algorithm to optimize the Kriging parameters and adopts in parallel the Compute Unified Device Architecture (CUDA) to run the PSO algorithm on the GPU side in order to raise the interpolated efficiency. Furthermore, a layer-constrained triangulated irregular network algorithm was proposed to construct the 3D geohazard bodies and the space geometry method was used to compute their volume information. The study also presents an implementation system to demonstrate the application of the framework and its related algorithms. This system makes a significant contribution to the demonstration and understanding of underground geohazard recognition in a three-dimensional environment.
|
|
| 36. |
|
|
| 37. |
- Wang, Zhenwu, et al.
(författare)
-
Interpreting convolutional neural network by joint evaluation of multiple feature maps and an improved NSGA-II algorithm
- 2024
-
Ingår i: Expert systems with applications. - 0957-4174 .- 1873-6793. ; 255
-
Tidskriftsartikel (refereegranskat)abstract
- The ’black box’ characteristics of Convolutional Neural Networks (CNNs) present significant risks to their application scenarios, such as reliability, security, and division of responsibilities. Addressing the interpretability of CNN emerges as an urgent and critical issue in the field of machine learning. Recent research on CNN interpretability has either yielded unstable or inconsistent interpretations, or produced coarse-scale interpretable heatmaps, limiting their applicability in various scenarios. In this work, we propose a novel method of CNNs interpretation by incorporating a joint evaluation of multiple feature maps and employing multi-objective optimization (JE&MOO-CAM). Firstly, a method of joint evaluation for all feature maps is proposed to preserve the complete object instances and improve the overall activation values. Secondly, an interpretation method of CNNs under the MOO framework is proposed to avoid the instability and inconsistency of interpretation. Finally, the operators of selection, crossover, and mutation, along with the method of population initialization in NSGA-II, are redesigned to properly express the characteristics of CNNs. The experimental results, including both qualitative and quantitative assessments along with a sanity check conducted on three classic CNN models—VGG16, AlexNet, and ResNet50—demonstrate the superior performance of the proposed JE&MOO-CAM model. This model not only accurately pinpoints the instances within the image requiring explanation but also preserves the integrity of these instances to the greatest extent possible. These capabilities signify that JE&MOO-CAM surpasses six other leading state-of-the-art methods across four established evaluation criteria.
|
|
| 38. |
- Wang, Zhenwu, et al.
(författare)
-
PML-ED : A method of partial multi-label learning by using encoder-decoder framework and exploring label correlation
- 2024
-
Ingår i: Information Sciences. - 0020-0255 .- 1872-6291. ; 661
-
Tidskriftsartikel (refereegranskat)abstract
- Partial multi-label learning (PML) addresses problems where each instance is assigned a candidate label set and only a subset of these candidate labels is correct. The major challenge of PML is that the training procedure can be easily misguided by noisy labels. Current studies on PML have revealed two significant drawbacks. First, most of them do not sufficiently explore complex label correlations, which could improve the effectiveness of label disambiguation. Second, PML models heavily rely on prior assumptions, limiting their applicability to specific scenarios. In this work, we propose a novel method of PML based on the Encoder-Decoder Framework (PML-ED) to address the drawbacks. PML-ED initially achieves the distribution of label probability through a KNN label attention mechanism. It then adopts Conditional Layer Normalization (CLN) to extract the high-order label correlation and relaxes the prior assumption of label noise by introducing a universal Encoder-Decoder framework. This approach makes PML-ED not only more efficient compared to the state-of-the-art methods, but also capable of handling the data with large noisy labels across different domains. Experimental results on 28 benchmark datasets demonstrate that the proposed PML-ED model, when benchmarked against nine leading-edge PML algorithms, achieves the highest average ranking across five evaluation criteria.
|
|
| 39. |
- Wang, Zhenwu, et al.
(författare)
-
Solving dynamic multi-objective optimization problems via quantifying intensity of environment changes and ensemble learning-based prediction strategies
- 2024
-
Ingår i: Applied Soft Computing. - 1568-4946 .- 1872-9681. ; 154
-
Tidskriftsartikel (refereegranskat)abstract
- Algorithms designed to solve dynamic multi-objective optimization problems (DMOPs) need to consider all of themultiple conflicting objectives to determine the optimal solutions. However, objective functions, constraints orparameters can change over time, which presents a considerable challenge. Algorithms should be able not only toidentify the optimal solution but also to quickly detect and respond to any changes of environment. In order toenhance the capability of detection and response to environmental changes, we propose a dynamic multiobjectiveoptimization (DMOO) algorithm based on the detection of environment change intensity andensemble learning (DMOO-DECI&EL). First, we propose a method for detecting environmental change intensity,where the change intensity is quantified and used to design response strategies. Second, a series of responsestrategies under the framework of ensemble learning are given to handle complex environmental changes.Finally, a boundary learning method is introduced to enhance the diversity and uniformity of the solutions.Experimental results on 14 benchmark functions demonstrate that the proposed DMOO-DECI&EL algorithmachieves the best comprehensive performance across three evaluation criteria, which indicates that DMOODECI&EL has better robustness and convergence and can generate solutions with better diversity compared tofive other state-of-the-art dynamic prediction strategies. In addition, the application of DMOO-DECI&EL to thereal-world scenario, namely the economic power dispatch problem, shows that the proposed method caneffectively handle real-world DMOPs.
|
|
| 40. |
- Wei, Yixuan, et al.
(författare)
-
Prediction of occupancy level and energy consumption in office building using blind system identification and neural networks
- 2019
-
Ingår i: Applied Energy. - : Elsevier. - 0306-2619 .- 1872-9118. ; 240, s. 276-294
-
Tidskriftsartikel (refereegranskat)abstract
- Occupancy behaviour plays an important role in energy consumption in buildings. Currently, the shallow understanding of occupancy has led to a considerable performance gap between predicted and measured energy use. This paper presents an approach to estimate the occupancy based on blind system identification (BSI), and a prediction model of electricity consumption by an air-conditioning system is developed and reported based on an artificial neural network with the BSI estimation of the number of occupants as an input. This starts from the identification of indoor CO2 dynamics derived from the mass-conservation law and venting levels. The unknown parameters, including the occupancy and model parameters, are estimated by using a frequentist maximum-likelihood algorithm and Bayesian estimation. The second phase is to establish the prediction model of the electricity consumption of the air-conditioning system by using a feed-forward neural network (FFNN) and extreme learning machine (ELM), as well as ensemble models. To analyse some aspects of the benchmark test for identifying the effect of structure parameters and input-selection alternatives, three studies are conducted on (1) the effect of predictor selection based on principal component analysis, (2) the effect of the estimated occupancy as the supplementary input, and (3) the effect of the neural network ensemble. The result shows that the occupancy number, as the input, is able to improve the accuracy in predicting energy consumption using a neural network model.
|
|
| 41. |
|
|
| 42. |
- Zhang, Xingxing, et al.
(författare)
-
ChatGPT for Fast Learning of Positive Energy District (PED) : A Trial Testing and Comparison with Expert Discussion Results
- 2023
-
Ingår i: Buildings. - 2075-5309. ; 13:6
-
Tidskriftsartikel (refereegranskat)abstract
- Positive energy districts (PEDs) are urban areas which seek to take an integral approach to climate neutrality by including technological, spatial, regulatory, financial, legal, social, and economic perspectives. It is still a new concept and approach for many stakeholders. ChatGPT, a generative pre-trained transformer, is an advanced artificial intelligence (AI) chatbot based on a complex network structure and trained by the company OpenAI. It has the potential for the fast learning of PED. This paper reports a trial test in which ChatGPT is used to provide written formulations of PEDs within three frameworks: challenge, impact, and communication and dissemination. The results are compared with the formulations derived from over 80 PED experts who took part in a two-day workshop discussing many aspects of PED research and development. The proposed methodology involves querying ChatGPT with specific questions and recording its responses. Subsequently, expert opinions on the same questions are provided to ChatGPT, aiming to elicit a comparison between the two sources of information. This approach enables an evaluation of ChatGPT’s answers in relation to the insights shared by domain experts. By juxtaposing the outputs, a comprehensive assessment can be made regarding the reliability, accuracy, and alignment of ChatGPT’s responses with expert viewpoints. It is found that ChatGPT can be a useful tool for the rapid formulation of basic information about PEDs that could be used for its wider dissemination amongst the general public. The model is also noted as having a number of limitations, such as providing pre-set single answers, a sensitivity to the phrasing of questions, a tendency to repeat non-important (or general) information, and an inability to assess inputs negatively or provide diverse answers to context-based questions. Its answers were not always based on up-to-date information. Other limitations and some of the ethical–social issues related to the use of ChatGPT are also discussed. This study not only validated the possibility of using ChatGPT to rapid study PEDs but also trained ChatGPT by feeding back the experts’ discussion into the tool. It is recommended that ChatGPT can be involved in real-time PED meetings or workshops so that it can be trained both iteratively and dynamically. © 2023 by the authors.
|
|
| 43. |
- Zhang, Xingxing, et al.
(författare)
-
Digital Twin for Accelerating Sustainability in Positive Energy District : A Review of Simulation Tools and Applications
- 2021
-
Ingår i: Frontiers in Sustainable Cities. - : Frontiers Media SA. - 2624-9634. ; 3
-
Forskningsöversikt (refereegranskat)abstract
- A digital twin is regarded as a potential solution to optimize positive energy districts (PED). This paper presents a compact review about digital twins for PED from aspects of concepts, working principles, tools/platforms, and applications, in order to address the issues of both how a digital PED twin is made and what tools can be used for a digital PED twin. Four key components of digital PED twin are identified, i.e., a virtual model, sensor network integration, data analytics, and a stakeholder layer. Very few available tools now have full functions for digital PED twin, while most tools either have a focus on industrial applications or are designed for data collection, communication and visualization based on building information models (BIM) or geographical information system (GIS). Several observations gained from successful application are that current digital PED twins can be categorized into three tiers: (1) an enhanced version of BIM model only, (2) semantic platforms for data flow, and (3) big data analysis and feedback operation. Further challenges and opportunities are found in areas of data analysis and semantic interoperability, business models, data security, and management. The outcome of the review is expected to provide useful information for further development of digital PED twins and optimizing its sustainability.
|
|
| 44. |
- Zhao, Jing, et al.
(författare)
-
Autoregressive count data modeling on mobility patterns to predict cases of COVID-19 infection
- 2022
-
Ingår i: Stochastic environmental research and risk assessment (Print). - : Springer Science and Business Media LLC. - 1436-3240 .- 1436-3259. ; 36:12, s. 4185-4200
-
Tidskriftsartikel (refereegranskat)abstract
- At the beginning of 2022 the global daily count of new cases of COVID-19 exceeded 3.2 million, a tripling of the historical peak value reported between the initial outbreak of the pandemic and the end of 2021. Aerosol transmission through interpersonal contact is the main cause of the disease's spread, although control measures have been put in place to reduce contact opportunities. Mobility pattern is a basic mechanism for understanding how people gather at a location and how long they stay there. Due to the inherent dependencies in disease transmission, models for associating mobility data with confirmed cases need to be individually designed for different regions and time periods. In this paper, we propose an autoregressive count data model under the framework of a generalized linear model to illustrate a process of model specification and selection. By evaluating a 14-day-ahead prediction from Sweden, the results showed that for a dense population region, using mobility data with a lag of 8 days is the most reliable way of predicting the number of confirmed cases in relative numbers at a high coverage rate. It is sufficient for both of the autoregressive terms, studied variable and conditional expectation, to take one day back. For sparsely populated regions, a lag of 10 days produced the lowest error in absolute value for the predictions, where weekly periodicity on the studied variable is recommended for use. Interventions were further included to identify the most relevant mobility categories. Statistical features were also presented to verify the model assumptions.
|
|
| 45. |
- Zhenwu, Wang, et al.
(författare)
-
Partial Classifier Chains with Feature Selection by Exploiting Label Correlation in Multi-Label Classification
- 2020
-
Ingår i: Entropy. - : MDPI. - 1099-4300. ; 22:1143, s. 1-22
-
Tidskriftsartikel (refereegranskat)abstract
- Multi-label classification (MLC) is a supervised learning problem where an object is naturally associated with multiple concepts because it can be described from various dimensions. How to exploit the resulting label correlations is the key issue in MLC problems. The classifier chain (CC) is a well-known MLC approach that can learn complex coupling relationships between labels. CC suffers from two obvious drawbacks: (1) label ordering is decided at random although it usually has a strong effect on predictive performance; (2) all the labels are inserted into the chain, although some of them may carry irrelevant information that discriminates against the others. In this work, we propose a partial classifier chain method with feature selection (PCC-FS) that exploits the label correlation between label and feature spaces and thus solves the two previously mentioned problems simultaneously. In the PCC-FS algorithm, feature selection is performed by learning the covariance between feature set and label set, thus eliminating the irrelevant features that can diminish classification performance. Couplings in the label set are extracted, and the coupled labels of each label are inserted simultaneously into the chain structure to execute the training and prediction activities. The experimental results from five metrics demonstrate that, in comparison to eight state-of-the-art MLC algorithms, the proposed method is a significant improvement on existing multi-label classification.
|
|
