To process the large amounts of data industrial systems are producing nowadays, machine learning techniques have shown their usefulness in many applications. As the amounts of data being generated are getting huge, the need for machine learning methods which can deal with them in an appropriate way -- i.e. methods which can be adapted incrementally -- becomes very important. Ensembles of classifiers have been shown to be able to improve the predictive accuracy as well as the robustness of single classification methods. In this work novel incremental variants of several well-known classifier fusion methods (Fuzzy Integral, Decision Templates, Dempster-Shafer Combination and Discounted Dempster-Shafer Combination) are presented. Furthermore, a novel incremental classifier fusion method called Incremental Direct Cluster-based fusion will be introduced, which exploits an evolving clustering approach. A flexible and interactive framework for on-line learning will be introduced, in which the ensemble (classifier fusion) methods are adapted incrementally in a sample-wise manner together with their base classifiers. The performance of this framework and the proposed incremental classifiers fusion methods therein are evaluated on five real-world visual quality inspection tasks, captured on-line from an industrial CD imprint production process.
{"title":"Increasing On-line Classification Performance Using Incremental Classifier Fusion","authors":"Davy Sannen, E. Lughofer, H. Brussel","doi":"10.1109/ICAIS.2009.25","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.25","url":null,"abstract":"To process the large amounts of data industrial systems are producing nowadays, machine learning techniques have shown their usefulness in many applications. As the amounts of data being generated are getting huge, the need for machine learning methods which can deal with them in an appropriate way -- i.e. methods which can be adapted incrementally -- becomes very important. Ensembles of classifiers have been shown to be able to improve the predictive accuracy as well as the robustness of single classification methods. In this work novel incremental variants of several well-known classifier fusion methods (Fuzzy Integral, Decision Templates, Dempster-Shafer Combination and Discounted Dempster-Shafer Combination) are presented. Furthermore, a novel incremental classifier fusion method called Incremental Direct Cluster-based fusion will be introduced, which exploits an evolving clustering approach. A flexible and interactive framework for on-line learning will be introduced, in which the ensemble (classifier fusion) methods are adapted incrementally in a sample-wise manner together with their base classifiers. The performance of this framework and the proposed incremental classifiers fusion methods therein are evaluated on five real-world visual quality inspection tasks, captured on-line from an industrial CD imprint production process.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"112 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122742855","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 purpose of this paper is twofold. On one hand, we propose a modification of the general Model Predictive Control (MPC) approach where a prespecified tracking error is tolerated. The introduction of error tolerance (ET) in the MPC optimization algorithm reduces considerably the average duration of each optimization step and makes the MPC computationally more efficient and attractive for industrial applications. On the other hand a challenging scheduled crystallization process serves as a case study to show the practical relevance of the new intelligent predictive control. Comparative tests with different control policies are performed: i) Classical MPC with analytical or Artificial Neural Network (ANN) process model; ii) ET MPC with analytical or ANN process model; iii) Proportional-Integral (PI) control. Besides the computational benefits of ET MPC, the integration of ANN into the ET MPC brings substantial improvements of the final process performance measures and further relaxes the computational demands.
本文的目的是双重的。一方面，我们提出了一种通用模型预测控制(MPC)方法的修改，其中允许预先指定的跟踪误差。在MPC优化算法中引入容错(ET)大大减少了每个优化步骤的平均持续时间，使MPC计算效率更高，对工业应用更有吸引力。另一方面，一个具有挑战性的计划结晶过程作为一个案例研究，展示了新的智能预测控制的实际意义。在不同控制策略下进行了对比试验:i)经典MPC与分析模型或人工神经网络(ANN)过程模型;ii) ET MPC与分析或人工神经网络过程模型;比例积分(PI)控制。除了ET MPC的计算效益外，将人工神经网络集成到ET MPC中，可以大大改善最终的过程性能指标，并进一步降低计算需求。
{"title":"Intelligent Predictive Control - Application to Scheduled Crystallization Processes","authors":"L. A. P. Suárez, P. Georgieva, S. Azevedo","doi":"10.1109/ICAIS.2009.34","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.34","url":null,"abstract":"The purpose of this paper is twofold. On one hand, we propose a modification of the general Model Predictive Control (MPC) approach where a prespecified tracking error is tolerated. The introduction of error tolerance (ET) in the MPC optimization algorithm reduces considerably the average duration of each optimization step and makes the MPC computationally more efficient and attractive for industrial applications. On the other hand a challenging scheduled crystallization process serves as a case study to show the practical relevance of the new intelligent predictive control. Comparative tests with different control policies are performed: i) Classical MPC with analytical or Artificial Neural Network (ANN) process model; ii) ET MPC with analytical or ANN process model; iii) Proportional-Integral (PI) control. Besides the computational benefits of ET MPC, the integration of ANN into the ET MPC brings substantial improvements of the final process performance measures and further relaxes the computational demands.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129619039","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}
In this paper we discuss the need of a novel control architecture for managing the growing complexity in modern vehicles and outline a multi-layered approach for self-management in adaptive automotive systems. With this multi-layered control architecture it is possible to react in an adequate and quick way to changes in the supervised technical system. Especially for complex distributed real-time systems with various different requirements and system objectives, like vehicles, this approach provides the necessary degree of flexibility and dependability. In a first evaluation of this control architecture in a realistic automotive scenario we show the advantages of the multi-layered approach compared to a traditional central control architecture.
{"title":"A Multi-layered Control Architecture for Self-Management in Adaptive Automotive Systems","authors":"M. Zeller, Gereon Weiss, D. Eilers, R. Knorr","doi":"10.1109/ICAIS.2009.20","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.20","url":null,"abstract":"In this paper we discuss the need of a novel control architecture for managing the growing complexity in modern vehicles and outline a multi-layered approach for self-management in adaptive automotive systems. With this multi-layered control architecture it is possible to react in an adequate and quick way to changes in the supervised technical system. Especially for complex distributed real-time systems with various different requirements and system objectives, like vehicles, this approach provides the necessary degree of flexibility and dependability. In a first evaluation of this control architecture in a realistic automotive scenario we show the advantages of the multi-layered approach compared to a traditional central control architecture.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"2014 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128923651","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 article discusses essential for systems adaptation issues. The investigation objectives are to analyse and compare abilities for self-regulation and adaptation of heuristic algorithm called Free Search. It is evaluated with hard constraint test problem. Experimental results are compared with collection of published in the literature solutions achieved by other methods.
{"title":"Free Search - A Model of Adaptive Intelligence","authors":"K. Penev","doi":"10.1109/ICAIS.2009.24","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.24","url":null,"abstract":"The article discusses essential for systems adaptation issues. The investigation objectives are to analyse and compare abilities for self-regulation and adaptation of heuristic algorithm called Free Search. It is evaluated with hard constraint test problem. Experimental results are compared with collection of published in the literature solutions achieved by other methods.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134241294","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}
A neuro-fuzzy adaptive control approach for nonlinear systems with model uncertainties is proposed. The implemented control scheme consists of a proportional plus derivative controller that is provided both to guarantee global asymptotic stability in compact space and as an inverse reference model of the response of the controlled system. Its output is used as an error signal by an on-line learning algorithm to update the parameters of a neuro-fuzzy feedback controller. The latter is able to gradually replace the conventional controller from the control of the system. The proposed learning algorithm makes direct use of the variable structure systems theory and establishes a sliding motion in terms of the neuro-fuzzy controller parameters. An integrating term has been additionally applied to the overall control signal of the two controllers and the performance of the control scheme has been tested on the wheel slip control problem within an antilock breaking system model. The analytical claims have been justified under the existence of model uncertainties and large initial errors.
{"title":"Neuro-Fuzzy Control of Antilock Braking System Using Variable-Structure-Systems-Based Learning Algorithm","authors":"A. Topalov, E. Kayacan, Y. Oniz, O. Kaynak","doi":"10.1109/ICAIS.2009.35","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.35","url":null,"abstract":"A neuro-fuzzy adaptive control approach for nonlinear systems with model uncertainties is proposed. The implemented control scheme consists of a proportional plus derivative controller that is provided both to guarantee global asymptotic stability in compact space and as an inverse reference model of the response of the controlled system. Its output is used as an error signal by an on-line learning algorithm to update the parameters of a neuro-fuzzy feedback controller. The latter is able to gradually replace the conventional controller from the control of the system. The proposed learning algorithm makes direct use of the variable structure systems theory and establishes a sliding motion in terms of the neuro-fuzzy controller parameters. An integrating term has been additionally applied to the overall control signal of the two controllers and the performance of the control scheme has been tested on the wheel slip control problem within an antilock breaking system model. The analytical claims have been justified under the existence of model uncertainties and large initial errors.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"115 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134252290","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 control of parameters during the execution of bio-inspired algorithms is an open research area. In this paper, we propose a new parameter control strategy for the immune algorithm CLONALG. Our approach is based on reinforcement learning ideas. We focus our attention on controlling the number of clones and the number of selected cells which follow a mutation process for improvement. Their values allow a trade-off between intensification and diversification of the search. Our approach provides an efficient and low cost adaptive technique for parameter control. We use instances of the Travelling Salesman Problem that has been tackled before by using CLONALG. The results obtained are very encouraging.
{"title":"A Dynamic Adaptive Calibration of the CLONALG Immune Algorithm","authors":"M. Riff, Elizabeth Montero","doi":"10.1109/ICAIS.2009.38","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.38","url":null,"abstract":"The control of parameters during the execution of bio-inspired algorithms is an open research area. In this paper, we propose a new parameter control strategy for the immune algorithm CLONALG. Our approach is based on reinforcement learning ideas. We focus our attention on controlling the number of clones and the number of selected cells which follow a mutation process for improvement. Their values allow a trade-off between intensification and diversification of the search. Our approach provides an efficient and low cost adaptive technique for parameter control. We use instances of the Travelling Salesman Problem that has been tackled before by using CLONALG. The results obtained are very encouraging.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134267446","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}
In this paper a self-adapting architecture for FOREX market prediction, which is being developed, is described. The proposed system utilizes genetic programming (GP) for predictor representation. The goal of the system is the design and adaptation of simple predictors which can either be used by the system itself or be 'manually' used by a human trader.
{"title":"Towards the Automatic Evolutionary Prediction of the FOREX Market Behaviour","authors":"Karel Slaný","doi":"10.1109/ICAIS.2009.31","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.31","url":null,"abstract":"In this paper a self-adapting architecture for FOREX market prediction, which is being developed, is described. The proposed system utilizes genetic programming (GP) for predictor representation. The goal of the system is the design and adaptation of simple predictors which can either be used by the system itself or be 'manually' used by a human trader.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124358228","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}
In this paper at first a new measure of stability of clustering solutions over different bootstrap samples of a data set is proposed. Thereafter in this paper, a multiobjective optimization based clustering technique is developed which optimizes both the measures of symmetry and stability simultaneously to automatically determine the appropriate number of clusters and the appropriate partitioning from data sets having symmetrical shaped clusters. The proposed algorithm utilizes a recently developed simulated annealing based multiobjective optimization technique, AMOSA, as the underlying optimization method. Here assignment of points to different clusters are done based on a recently developed point symmetry based distance rather than the Euclidean distance. Results on several artificial and real-life data sets show that the proposed technique is well-suited to detect the number of clusters from data sets having point symmetric clusters.
{"title":"A New Measure of Stability of Clustering Solutions: Application to Data Partitioning","authors":"S. Saha, S. Bandyopadhyay","doi":"10.1109/ICAIS.2009.37","DOIUrl":"https://doi.org/10.1109/ICAIS.2009.37","url":null,"abstract":"In this paper at first a new measure of stability of clustering solutions over different bootstrap samples of a data set is proposed. Thereafter in this paper, a multiobjective optimization based clustering technique is developed which optimizes both the measures of symmetry and stability simultaneously to automatically determine the appropriate number of clusters and the appropriate partitioning from data sets having symmetrical shaped clusters. The proposed algorithm utilizes a recently developed simulated annealing based multiobjective optimization technique, AMOSA, as the underlying optimization method. Here assignment of points to different clusters are done based on a recently developed point symmetry based distance rather than the Euclidean distance. Results on several artificial and real-life data sets show that the proposed technique is well-suited to detect the number of clusters from data sets having point symmetric clusters.","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"92 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127325770","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 : 1900-01-01DOI: 10.1007/978-3-642-16236-7_1
M. Gouko, Koji Ito
{"title":"Environmental Modeling and Identification Based on Changes in Sensory Information","authors":"M. Gouko, Koji Ito","doi":"10.1007/978-3-642-16236-7_1","DOIUrl":"https://doi.org/10.1007/978-3-642-16236-7_1","url":null,"abstract":"","PeriodicalId":161840,"journal":{"name":"2009 International Conference on Adaptive and Intelligent Systems","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116368024","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：