Pub Date : 2022-12-15DOI: 10.1109/ICNSC55942.2022.10004160
Jiale Yang, Han Yang, Fei Wang, Xiong-Zi Chen
Object detection in UAV image processing has gradually become a hot research topic in recent years. The performance of general object detection algorithms tends to degrade significantly when applied to UAV scenes. This is due to the fact that UAV images are taken from high altitude with high resolution and a large proportion of small objects. In order to improve the precision of UAV object detection while satisfying the lightweight feature, we modify the YOLOv5s model. To address the small object detection problem, a prediction head is added to better retain small object feature information. The CBAM attention module is also integrated to better find attention regions in dense scenes. The original IOU-NMS is replaced by NWD-NMS in post-processing to alleviate the sensitivity of IOU to small objects. Experiments show that our method has good performance on the dataset Visdrone-2020, and the mAP is significantly improved from the original.
{"title":"A modified YOLOv5 for object detection in UAV-captured scenarios","authors":"Jiale Yang, Han Yang, Fei Wang, Xiong-Zi Chen","doi":"10.1109/ICNSC55942.2022.10004160","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004160","url":null,"abstract":"Object detection in UAV image processing has gradually become a hot research topic in recent years. The performance of general object detection algorithms tends to degrade significantly when applied to UAV scenes. This is due to the fact that UAV images are taken from high altitude with high resolution and a large proportion of small objects. In order to improve the precision of UAV object detection while satisfying the lightweight feature, we modify the YOLOv5s model. To address the small object detection problem, a prediction head is added to better retain small object feature information. The CBAM attention module is also integrated to better find attention regions in dense scenes. The original IOU-NMS is replaced by NWD-NMS in post-processing to alleviate the sensitivity of IOU to small objects. Experiments show that our method has good performance on the dataset Visdrone-2020, and the mAP is significantly improved from the original.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117016766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Feature selection is an essential technique which has been widely applied in data mining. Recent research has shown that a good feature subset can be obtained by using evolutionary computing (EC) approaches as a wrapper. However, most feature selection methods based on EC use a fixed-length encoding to represent feature subsets. When this fixed length representation is applied to high-dimensional data, it requires a large amount of memory space as well as a high computational cost. Moreover, this representation is inflexible and may limit the performance of EC because of a too huge search space. In this paper, we propose an Adaptive- Variable-Length Genetic Algorithm (A VLGA), which adopts a variable-length individual encoding and enables individuals with different lengths in a population to evolve in their own search space. An adaptive length changing mechanism is introduced which can extend or shorten an individual to guide it to explore in a better search space. Thus, A VLGA is able to adaptively concentrate on a smaller but more fruitful search space and yield better solutions more quickly. Experimental results on 6 high-dimensional datasets reveal that A VLGA performs significantly better than existing methods.
{"title":"High-dimensional Feature Selection in Classification: A Length-Adaptive Evolutionary Approach","authors":"Junhai Zhou, Jian-chun Lu, Quanwang Wu, Junhao Wen","doi":"10.1109/ICNSC55942.2022.10004048","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004048","url":null,"abstract":"Feature selection is an essential technique which has been widely applied in data mining. Recent research has shown that a good feature subset can be obtained by using evolutionary computing (EC) approaches as a wrapper. However, most feature selection methods based on EC use a fixed-length encoding to represent feature subsets. When this fixed length representation is applied to high-dimensional data, it requires a large amount of memory space as well as a high computational cost. Moreover, this representation is inflexible and may limit the performance of EC because of a too huge search space. In this paper, we propose an Adaptive- Variable-Length Genetic Algorithm (A VLGA), which adopts a variable-length individual encoding and enables individuals with different lengths in a population to evolve in their own search space. An adaptive length changing mechanism is introduced which can extend or shorten an individual to guide it to explore in a better search space. Thus, A VLGA is able to adaptively concentrate on a smaller but more fruitful search space and yield better solutions more quickly. Experimental results on 6 high-dimensional datasets reveal that A VLGA performs significantly better than existing methods.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114151267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-15DOI: 10.1109/ICNSC55942.2022.10004167
Haoran Gao, Zhijun Ding
Most research on machine learning relies on the availability of ground truth labels immediately after prediction. However, in many cases, the ground truth labels become available with a non-negligible delay. Considering that there is a large amount of unlabeled data in delayed labels, supervised model cannot utilize unlabeled data. Therefore, most of the research on delayed labels begins to train semi-supervised models in delayed labels. However, most research on delayed labels ignores that the labels of unlabeled data will arrive after several periods in delayed labels. Neither supervised nor semi-supervised models can solve the problem in delayed labels effectively. Besides, there remains a problem of concept drift due to the long period of data. In this paper, we propose an incremental learning model that can adapt to delayed labels. First, we should detect whether the concept drift takes place. Then we use knowledge distillation to update supervised and semi-supervised models while retaining the corresponding knowledge of past labeled data. Finally, we combine the supervised and semi-supervised models to make predictions. Finally, we apply our algorithms to synthetic and real credit scoring datasets. The experiment results indicate our algorithms have superiority in delayed labels.
{"title":"A Novel Machine Learning Method for Delayed Labels","authors":"Haoran Gao, Zhijun Ding","doi":"10.1109/ICNSC55942.2022.10004167","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004167","url":null,"abstract":"Most research on machine learning relies on the availability of ground truth labels immediately after prediction. However, in many cases, the ground truth labels become available with a non-negligible delay. Considering that there is a large amount of unlabeled data in delayed labels, supervised model cannot utilize unlabeled data. Therefore, most of the research on delayed labels begins to train semi-supervised models in delayed labels. However, most research on delayed labels ignores that the labels of unlabeled data will arrive after several periods in delayed labels. Neither supervised nor semi-supervised models can solve the problem in delayed labels effectively. Besides, there remains a problem of concept drift due to the long period of data. In this paper, we propose an incremental learning model that can adapt to delayed labels. First, we should detect whether the concept drift takes place. Then we use knowledge distillation to update supervised and semi-supervised models while retaining the corresponding knowledge of past labeled data. Finally, we combine the supervised and semi-supervised models to make predictions. Finally, we apply our algorithms to synthetic and real credit scoring datasets. The experiment results indicate our algorithms have superiority in delayed labels.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125974912","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The spherical search algorithm (SS) is a novel and competitive algorithm applied to real-world problems. However, the population of SS algorithm is divided equally, which requires a large number of computation resources for different problems. To alleviate the issues, we propose an immigration strategy-based spherical search algorithm, namely ISS. ISS adaptively selects individuals that are successfully updated in each generation and replaces the operator in the next iteration. The experiments were conducted on the 30 benchmark functions from the IEEE CEC2017. ISS is compared with SS to verify the effectiveness of the proposed adaptive immigration strategy. Additionally, the classical differential evolutionary algorithm (DE) and a state-of-the-art triple archive particle swarm optimization (TAPSO) are compared to test its performance further. The population diversity is analyzed to discuss the effect of ISS. The experimental results demonstrate that the proposed immigration strategy is quite effective, and ISS is significantly better than its peer's algorithms.
{"title":"An Immigration Strategy-based Spherical Search Algorithm","authors":"Qingya Sui, Sichen Tao, Lin Zhong, Haichuan Yang, Zhenyu Lei, Shangce Gao","doi":"10.1109/ICNSC55942.2022.10004149","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004149","url":null,"abstract":"The spherical search algorithm (SS) is a novel and competitive algorithm applied to real-world problems. However, the population of SS algorithm is divided equally, which requires a large number of computation resources for different problems. To alleviate the issues, we propose an immigration strategy-based spherical search algorithm, namely ISS. ISS adaptively selects individuals that are successfully updated in each generation and replaces the operator in the next iteration. The experiments were conducted on the 30 benchmark functions from the IEEE CEC2017. ISS is compared with SS to verify the effectiveness of the proposed adaptive immigration strategy. Additionally, the classical differential evolutionary algorithm (DE) and a state-of-the-art triple archive particle swarm optimization (TAPSO) are compared to test its performance further. The population diversity is analyzed to discuss the effect of ISS. The experimental results demonstrate that the proposed immigration strategy is quite effective, and ISS is significantly better than its peer's algorithms.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131717007","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-15DOI: 10.1109/ICNSC55942.2022.10004071
Peiyun Zhang, YanHao Tao, Junliang Shu
In a blockchain network, the instability of the block transmission process can affect the speed of block transmission. If blocks cannot be accepted by nodes and saved on a blockchain in time, which may lead to inconsistent blockchain ledgers stored by nodes, thus reducing the security of blockchain networks. However, when nodes transmit blocks, they often encounter problems of too large blocks and insufficient bandwidth, which results in slow block transmission speed and low efficiency. To solve the problems, it proposes a block transmission model, which encodes units into packets. Based on the model, the corresponding encoding and decoding processes are designed. The proposed method is compared with two state-of-the-art methods: Velocity and Kadcast. Experimental results show that the proposed method performs better than its peers in terms of block synchronization time, block transmission success ratio, and packet retransmission ratio.
{"title":"A Novel Block Transmission Model in Blockchain Networks","authors":"Peiyun Zhang, YanHao Tao, Junliang Shu","doi":"10.1109/ICNSC55942.2022.10004071","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004071","url":null,"abstract":"In a blockchain network, the instability of the block transmission process can affect the speed of block transmission. If blocks cannot be accepted by nodes and saved on a blockchain in time, which may lead to inconsistent blockchain ledgers stored by nodes, thus reducing the security of blockchain networks. However, when nodes transmit blocks, they often encounter problems of too large blocks and insufficient bandwidth, which results in slow block transmission speed and low efficiency. To solve the problems, it proposes a block transmission model, which encodes units into packets. Based on the model, the corresponding encoding and decoding processes are designed. The proposed method is compared with two state-of-the-art methods: Velocity and Kadcast. Experimental results show that the proposed method performs better than its peers in terms of block synchronization time, block transmission success ratio, and packet retransmission ratio.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129799520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The recycling and reuse of waste products is increasingly inseparable from disassembly lines. This study addresses a multi-product hybrid disassembly line balancing problem that is composed of a U-shaped line and a single-row layout. Disabled workers are considered. According to the characteristics of the problem, a mathematical model for maximizing the recovery profit is established. Combined with the actual disassembly process and considering the limitations of the disabled workers in product selection, two types of neighborhood structures are designed using a parallel neighborhood search algorithm (PNSA) to find feasible solutions. Experimental analysis shows that the model and algorithm proposed in this paper effectively solve the above problems.
{"title":"A Parallel Neighborhood Search Algorithm for Hybrid Disassembly Line Balancing Problem Considering Disabled Workers","authors":"Zheng Dou, Peisheng Liu, Xiwang Guo, Jiacun Wang, Shujin Qin, Liang Qi","doi":"10.1109/ICNSC55942.2022.10004123","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004123","url":null,"abstract":"The recycling and reuse of waste products is increasingly inseparable from disassembly lines. This study addresses a multi-product hybrid disassembly line balancing problem that is composed of a U-shaped line and a single-row layout. Disabled workers are considered. According to the characteristics of the problem, a mathematical model for maximizing the recovery profit is established. Combined with the actual disassembly process and considering the limitations of the disabled workers in product selection, two types of neighborhood structures are designed using a parallel neighborhood search algorithm (PNSA) to find feasible solutions. Experimental analysis shows that the model and algorithm proposed in this paper effectively solve the above problems.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129109317","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-15DOI: 10.1109/ICNSC55942.2022.10004138
Qiang-Zhang, Wen-Feng Li, Jingyu Zhou
To improve the efficiency of container terminal transshipment, the process of dual AGV's no-load formation running to the handling point from the control point of view is studied in this paper. It focuses on the formation process stability and the accuracy of dual AGVs parking at the destination. In this study, the leader-follower formation strategy is used to calculate the desired position and posture of the follower AGV. The position and posture errors are analyzed based on the kinematics model of AGV with nonholonomic constraints. Moreover, the sliding mode controller is designed, which uses position and posture errors as the control parameters. Finally, linear and curvilinear conditions are used to test the comprehensive performance of the formation strategy and controller. Simulation results show that the designed controller achieves fast formation and stable formation kept of dual AGVs with different initial errors. Foremost, the high accuracy in position and posture of dual AGVs parking at the destination can shorten adaptation time between the spreader and AGVs, which proves the dual AGVs formation scheme and controller designed in this paper are feasible and effective.
{"title":"No-Load Formation Control of Dual AGVs Based on Container Terminals","authors":"Qiang-Zhang, Wen-Feng Li, Jingyu Zhou","doi":"10.1109/ICNSC55942.2022.10004138","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004138","url":null,"abstract":"To improve the efficiency of container terminal transshipment, the process of dual AGV's no-load formation running to the handling point from the control point of view is studied in this paper. It focuses on the formation process stability and the accuracy of dual AGVs parking at the destination. In this study, the leader-follower formation strategy is used to calculate the desired position and posture of the follower AGV. The position and posture errors are analyzed based on the kinematics model of AGV with nonholonomic constraints. Moreover, the sliding mode controller is designed, which uses position and posture errors as the control parameters. Finally, linear and curvilinear conditions are used to test the comprehensive performance of the formation strategy and controller. Simulation results show that the designed controller achieves fast formation and stable formation kept of dual AGVs with different initial errors. Foremost, the high accuracy in position and posture of dual AGVs parking at the destination can shorten adaptation time between the spreader and AGVs, which proves the dual AGVs formation scheme and controller designed in this paper are feasible and effective.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128231652","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-15DOI: 10.1109/ICNSC55942.2022.10004050
Chadi M'Sila, R. Ayad, N. A. Oufroukh
The presence of Foreign Object Debris (FOD) on airport platforms constitutes a big risk, both for aircraft and for personnel. This debris, whatever its nature or size, whether it's a private effect, a tool, a component from an aircraft, or any object, As soon because it isn't observed and removed, it's liable becoming a FOD within the moving area. FOD can even be violently projected by jet blast, which might cause damage to other aircraft and injure personnel on the bottom, This paper discuss briefly FOD detection systems and the use of unmanned aerial systems for an automated FOD detection system on runways, which involves taking images of the runway with an Unmanned Aerial Vehicle (UAV), which could be detected and identified using artificial intelligence techniques. The method for determining an exact FOD position from aerial data is described in this study using a perspective projection transformation is used to determine the object's location in the field. For accurate findings, a strong object detection is essential, which is why the cutting-edge deep neural network YOLOV5 is used with both DeepSort Object tracking method. The paper represent an Automated UAV Navigation with PID control based for path tracking. A GUI that has been developed alow the operator to select the runway's intended path to be scanned and visualize the tracked FOD that has been found and its position in order to send a report that the operator can erase from the runway. The proposed system was assessed in real-time testing and a built-in Simulation under GAZEBO using the commercial quad copter Bebop connected to a base station operating under the Robot Operating System (ROS). our approach successfully identified several FODs using a combination of YOLOv5 and deepsort with an inference speed of 30 fps with a high accuarcy over 80%. The advantages of this system is the fulfilment of the FAA performance criteria of an AFDS, it facilitate the FOD scanning operation by using a graphical user interface that allow the operator to start the FOD scanning operation by selecting only the interested area in the runway, drone navigation tests with a 10 m/s wind speed were satisfactory, as well as it's ability to locate and send report of the detected FODs with small distance error less than 40 cm while a drone navigate with a 5m/s speed.
{"title":"Automated Foreign Object Debris Detection System based on UAV","authors":"Chadi M'Sila, R. Ayad, N. A. Oufroukh","doi":"10.1109/ICNSC55942.2022.10004050","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004050","url":null,"abstract":"The presence of Foreign Object Debris (FOD) on airport platforms constitutes a big risk, both for aircraft and for personnel. This debris, whatever its nature or size, whether it's a private effect, a tool, a component from an aircraft, or any object, As soon because it isn't observed and removed, it's liable becoming a FOD within the moving area. FOD can even be violently projected by jet blast, which might cause damage to other aircraft and injure personnel on the bottom, This paper discuss briefly FOD detection systems and the use of unmanned aerial systems for an automated FOD detection system on runways, which involves taking images of the runway with an Unmanned Aerial Vehicle (UAV), which could be detected and identified using artificial intelligence techniques. The method for determining an exact FOD position from aerial data is described in this study using a perspective projection transformation is used to determine the object's location in the field. For accurate findings, a strong object detection is essential, which is why the cutting-edge deep neural network YOLOV5 is used with both DeepSort Object tracking method. The paper represent an Automated UAV Navigation with PID control based for path tracking. A GUI that has been developed alow the operator to select the runway's intended path to be scanned and visualize the tracked FOD that has been found and its position in order to send a report that the operator can erase from the runway. The proposed system was assessed in real-time testing and a built-in Simulation under GAZEBO using the commercial quad copter Bebop connected to a base station operating under the Robot Operating System (ROS). our approach successfully identified several FODs using a combination of YOLOv5 and deepsort with an inference speed of 30 fps with a high accuarcy over 80%. The advantages of this system is the fulfilment of the FAA performance criteria of an AFDS, it facilitate the FOD scanning operation by using a graphical user interface that allow the operator to start the FOD scanning operation by selecting only the interested area in the runway, drone navigation tests with a 10 m/s wind speed were satisfactory, as well as it's ability to locate and send report of the detected FODs with small distance error less than 40 cm while a drone navigate with a 5m/s speed.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115574651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the development of economy, residential power users account for a higher and higher proportion in the power system. The modern power system focusing on residential load needs to realize the stability of load demand changes by combining forecasting information with long and short term dispatching. However, residential micro grid load usually has high fluctuation, so it is a challenging problem to achieve accurate prediction. Based on the characteristics of residential power load, this paper studies the short-term forecasting task of residential power load. BILSTM-MDN hybrid prediction models were constructed by BiLSTM's ability to learn long-term dependence and underlying correlation logic. Finally, 50 apartment load data sets are used to verify the great potential of the model based on BiLSTM-MDN in residential short-term power load prediction with high fluctuation. The accuracy of prediction reached MAPE 18.25% and RMSE 30.53%.
{"title":"A Short-term Residential Load Forecast Model Based on BiLSTM-MDN","authors":"Rushan Zheng, Jian Yu, Yizhen Wang, Xiongbing Chen","doi":"10.1109/ICNSC55942.2022.10004172","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004172","url":null,"abstract":"With the development of economy, residential power users account for a higher and higher proportion in the power system. The modern power system focusing on residential load needs to realize the stability of load demand changes by combining forecasting information with long and short term dispatching. However, residential micro grid load usually has high fluctuation, so it is a challenging problem to achieve accurate prediction. Based on the characteristics of residential power load, this paper studies the short-term forecasting task of residential power load. BILSTM-MDN hybrid prediction models were constructed by BiLSTM's ability to learn long-term dependence and underlying correlation logic. Finally, 50 apartment load data sets are used to verify the great potential of the model based on BiLSTM-MDN in residential short-term power load prediction with high fluctuation. The accuracy of prediction reached MAPE 18.25% and RMSE 30.53%.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126683376","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-15DOI: 10.1109/ICNSC55942.2022.10004137
Xiaolin Wang, Q. Kang, Mengchu Zhou
Multi-task optimization (MTO) is a novel emerging evolutionary computation paradigm that is used for solving multiple optimization tasks concurrently. Most MTO algorithms limit each individual to one task, and thus weaken the performance of information exchange. To address this issue and improve the efficiency of knowledge transfer, this work proposes an efficient MTO framework named individually-guided multi-task optimization (IMTO). It divides evolutions into vertical and horizontal ones. To further improve the efficiency of knowledge transfer, a partial individuals' learning scheme is used to choose suitable individuals to learn from other tasks. Experimental results show its superior advantages over the multifactorial evolutionary algorithm and its variants.
{"title":"Individually-guided Evolutionary Algorithm for Solving Multi-task Optimization Problems","authors":"Xiaolin Wang, Q. Kang, Mengchu Zhou","doi":"10.1109/ICNSC55942.2022.10004137","DOIUrl":"https://doi.org/10.1109/ICNSC55942.2022.10004137","url":null,"abstract":"Multi-task optimization (MTO) is a novel emerging evolutionary computation paradigm that is used for solving multiple optimization tasks concurrently. Most MTO algorithms limit each individual to one task, and thus weaken the performance of information exchange. To address this issue and improve the efficiency of knowledge transfer, this work proposes an efficient MTO framework named individually-guided multi-task optimization (IMTO). It divides evolutions into vertical and horizontal ones. To further improve the efficiency of knowledge transfer, a partial individuals' learning scheme is used to choose suitable individuals to learn from other tasks. Experimental results show its superior advantages over the multifactorial evolutionary algorithm and its variants.","PeriodicalId":230499,"journal":{"name":"2022 IEEE International Conference on Networking, Sensing and Control (ICNSC)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125409551","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：