Pub Date : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931771
Willemijn Remmerswaal, D. Sun, A. Jamshidnejad, B. Schutter
In general, the performance of model-based controllers cannot be guaranteed under model uncertainties or disturbances, while learning-based controllers require an extensively sufficient training process to perform well. These issues especially hold for large-scale nonlinear systems such as urban traffic networks. In this paper, a new framework is proposed by combining model predictive control (MPC) and reinforcement learning (RL) to provide desired performance for urban traffic networks even during the learning process, despite model uncertainties and disturbances. MPC and RL complement each other very well, since MPC provides a sub-optimal and constraint-satisfying control input while RL provides adaptive control laws and can handle uncertainties and disturbances. The resulting combined framework is applied for traffic signal control (TSC) of an urban traffic network. A case study is carried out to compare the performance of the proposed framework and other baseline controllers. Results show that the proposed combined framework outperforms conventional control methods under system uncertainties, in terms of reducing traffic congestion.
{"title":"Combined MPC and reinforcement learning for traffic signal control in urban traffic networks","authors":"Willemijn Remmerswaal, D. Sun, A. Jamshidnejad, B. Schutter","doi":"10.1109/ICSTCC55426.2022.9931771","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931771","url":null,"abstract":"In general, the performance of model-based controllers cannot be guaranteed under model uncertainties or disturbances, while learning-based controllers require an extensively sufficient training process to perform well. These issues especially hold for large-scale nonlinear systems such as urban traffic networks. In this paper, a new framework is proposed by combining model predictive control (MPC) and reinforcement learning (RL) to provide desired performance for urban traffic networks even during the learning process, despite model uncertainties and disturbances. MPC and RL complement each other very well, since MPC provides a sub-optimal and constraint-satisfying control input while RL provides adaptive control laws and can handle uncertainties and disturbances. The resulting combined framework is applied for traffic signal control (TSC) of an urban traffic network. A case study is carried out to compare the performance of the proposed framework and other baseline controllers. Results show that the proposed combined framework outperforms conventional control methods under system uncertainties, in terms of reducing traffic congestion.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121414280","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931831
Ravi kiran Inapakurthi, K. Mitra
Nonlinear system identification of complex and nonlinear unit operations and unit processes requires accurate modelling approaches. For this, first-principles based models were initially explored as they enable the causal explanation available among variables. However, the numerical integration issues along with the availability of voluminous data for developing data-based models has resulted in the shift from the conventional modelling approach to Machine Learning (ML) based modelling. In this study, Support Vector Regression (SVR) is used to model complex Industrial Grinding Circuit (IGC). To aid the accurate model requirement in process systems engineering domain, the tunable parameters of SVR are optimized using a novel multi-objective optimization formulation, which helps in minimizing the chances of over-fitting while simultaneously ensuring accurate models for IGC. The formulation is optimized using evolutionary algorithm to track and retain the most accurate models. The Pareto optimal SVR models have a minimum accuracy of 99. 786% and the prediction performance of the best model selected using knee point from the Pareto optimal set is compared with a model selected using arbitrary approach to show the competitiveness of the proposed technique.
{"title":"System Identification and Process Modelling of Dynamic Systems Using Machine Learning","authors":"Ravi kiran Inapakurthi, K. Mitra","doi":"10.1109/ICSTCC55426.2022.9931831","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931831","url":null,"abstract":"Nonlinear system identification of complex and nonlinear unit operations and unit processes requires accurate modelling approaches. For this, first-principles based models were initially explored as they enable the causal explanation available among variables. However, the numerical integration issues along with the availability of voluminous data for developing data-based models has resulted in the shift from the conventional modelling approach to Machine Learning (ML) based modelling. In this study, Support Vector Regression (SVR) is used to model complex Industrial Grinding Circuit (IGC). To aid the accurate model requirement in process systems engineering domain, the tunable parameters of SVR are optimized using a novel multi-objective optimization formulation, which helps in minimizing the chances of over-fitting while simultaneously ensuring accurate models for IGC. The formulation is optimized using evolutionary algorithm to track and retain the most accurate models. The Pareto optimal SVR models have a minimum accuracy of 99. 786% and the prediction performance of the best model selected using knee point from the Pareto optimal set is compared with a model selected using arbitrary approach to show the competitiveness of the proposed technique.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"152 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123062034","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931801
Sayeh Rezaee, César Nieto, Abhyudai Singh
We consider the problem of transmitting the state value of a dynamical system through a communication network. The dynamics of the error in state estimation is modeled using a stochastic hybrid system formalism, where the error grows exponentially over time. Transmission occurs over the network at specific times to acquire the system's state, and whenever a transmission is triggered, the error is reset to a zero-mean random variable. Our goal is to uncover transmission strategies that minimize a combination of the steady-state error variance and the average number of transmissions per unit of time. We found that a constant Poisson rate of transmission results in a heavy-tailed distribution for the estimation error. Next, we consider a random non-threshold transmission rate that varies as a power law of the error. Finally, we explore a threshold-based rate in which transmission occurs exactly when the error reaches a threshold. Our results show that if the error's variance after transmission is small enough, a threshold-based strategy is the optimal paradigm. On the other hand, if this variance is large, and the error does not grow fast enough, the random non-threshold transmission strategy emerges as optimal. These analytical results are verified by simulations of the stochastic hybrid system.
{"title":"Optimal network transmission to minimize state-estimation error and channel usage","authors":"Sayeh Rezaee, César Nieto, Abhyudai Singh","doi":"10.1109/ICSTCC55426.2022.9931801","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931801","url":null,"abstract":"We consider the problem of transmitting the state value of a dynamical system through a communication network. The dynamics of the error in state estimation is modeled using a stochastic hybrid system formalism, where the error grows exponentially over time. Transmission occurs over the network at specific times to acquire the system's state, and whenever a transmission is triggered, the error is reset to a zero-mean random variable. Our goal is to uncover transmission strategies that minimize a combination of the steady-state error variance and the average number of transmissions per unit of time. We found that a constant Poisson rate of transmission results in a heavy-tailed distribution for the estimation error. Next, we consider a random non-threshold transmission rate that varies as a power law of the error. Finally, we explore a threshold-based rate in which transmission occurs exactly when the error reaches a threshold. Our results show that if the error's variance after transmission is small enough, a threshold-based strategy is the optimal paradigm. On the other hand, if this variance is large, and the error does not grow fast enough, the random non-threshold transmission strategy emerges as optimal. These analytical results are verified by simulations of the stochastic hybrid system.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115837328","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931774
Valentin Plamenov Chernev, Lino O. Santos, A. Wouwer, A. Kienle
Simulated moving bed chromatographic (SMB) processes are used for difficult separations in pharmaceutical, biotechnological and petrochemical industries. Due to high sensitivity to disturbances these processes are usually operated in open-loop mode under suboptimal conditions. In the present work, operation of such processes based on the online optimizing model predictive control (MPC) using the full blown chromatographic model is proposed. For the fast and accurate solution of the underlying model described by a system of partial differential algebraic equations, the so-called space-time conservation element/solution method (CE/SE) is used. As an application example, the separation of racemic mixture of bicalutamides, one of which is a valuable active pharmaceutical component, is considered. To evaluate the performance of the controller, reference tracking (change of the purity requirements) and disturbance rejection (change of the composition of the feed mixture) scenarios are simulated. Since there is no plant-model mismatch, the controller is able to follow the change of the reference from complete to reduced purity separation closely. However, the results of the disturbance rejection simulation shows that the controller requires an adaption mechanism in order to efficiently reject the disturbance.
{"title":"Model Predictive Control of Simulated Moving Bed Chromatographic Processes Using Conservation Element/Solution Element Method","authors":"Valentin Plamenov Chernev, Lino O. Santos, A. Wouwer, A. Kienle","doi":"10.1109/ICSTCC55426.2022.9931774","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931774","url":null,"abstract":"Simulated moving bed chromatographic (SMB) processes are used for difficult separations in pharmaceutical, biotechnological and petrochemical industries. Due to high sensitivity to disturbances these processes are usually operated in open-loop mode under suboptimal conditions. In the present work, operation of such processes based on the online optimizing model predictive control (MPC) using the full blown chromatographic model is proposed. For the fast and accurate solution of the underlying model described by a system of partial differential algebraic equations, the so-called space-time conservation element/solution method (CE/SE) is used. As an application example, the separation of racemic mixture of bicalutamides, one of which is a valuable active pharmaceutical component, is considered. To evaluate the performance of the controller, reference tracking (change of the purity requirements) and disturbance rejection (change of the composition of the feed mixture) scenarios are simulated. Since there is no plant-model mismatch, the controller is able to follow the change of the reference from complete to reduced purity separation closely. However, the results of the disturbance rejection simulation shows that the controller requires an adaption mechanism in order to efficiently reject the disturbance.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132480174","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931808
Ovidiu Pauca, A. Maxim, C. Caruntu
Travelling on crowded roads, vehicles usually move in a dynamic environment where, fixed and mobile obstacles can be detected. This paper proposes a solution for a trajectory planner that can be used to compute paths so that, both fixed and mobile obstacles can be overpassed. The trajectory planner, based on the information from sensors regarding the distance to the other vehicles, computes the time moments when the vehicle has to overpass the obstacles. The generated trajectory is a time function that, based on these time moments, leads the vehicle to avoid a collision with the obstacles. The proposed trajectory planner is tested in three scenarios, where the obstacles are fixed or mobile. The results prove the effectiveness of the proposed trajectory planning solution.
{"title":"Vehicle Trajectory Planning for Collision Avoidance with Mobile Obstacles","authors":"Ovidiu Pauca, A. Maxim, C. Caruntu","doi":"10.1109/ICSTCC55426.2022.9931808","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931808","url":null,"abstract":"Travelling on crowded roads, vehicles usually move in a dynamic environment where, fixed and mobile obstacles can be detected. This paper proposes a solution for a trajectory planner that can be used to compute paths so that, both fixed and mobile obstacles can be overpassed. The trajectory planner, based on the information from sensors regarding the distance to the other vehicles, computes the time moments when the vehicle has to overpass the obstacles. The generated trajectory is a time function that, based on these time moments, leads the vehicle to avoid a collision with the obstacles. The proposed trajectory planner is tested in three scenarios, where the obstacles are fixed or mobile. The results prove the effectiveness of the proposed trajectory planning solution.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"364 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132506260","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931794
Christian Utama, B. Karg, Christian Meske, S. Lucia
Model predictive control (MPC) has been established in a wide range of control applications as the standard approach. But applying MPC requires solving a potentially complex optimization problem online to generate a new control input signal. To avoid the expensive online computations, deep learning-based MPC has been developed, in which neural networks imitate the behavior of the MPC. When such a data-driven approximate controller is derived, there is no straightforward way to trace the reasons for its proposed actions back to its inputs, hence making the controller a black-box model. In this paper, we propose the use of SHAP, an explainable artifical intelligence technique, to generate insights from learning-based MPC for the purpose of model debugging and simplification. Our results show that SHAP can explain general control behaviors and can also support model simplification in an informed way, representing a better alternative to dimensionality reduction techniques such as principal component analysis.
{"title":"Explainable artificial intelligence for deep learning-based model predictive controllers","authors":"Christian Utama, B. Karg, Christian Meske, S. Lucia","doi":"10.1109/ICSTCC55426.2022.9931794","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931794","url":null,"abstract":"Model predictive control (MPC) has been established in a wide range of control applications as the standard approach. But applying MPC requires solving a potentially complex optimization problem online to generate a new control input signal. To avoid the expensive online computations, deep learning-based MPC has been developed, in which neural networks imitate the behavior of the MPC. When such a data-driven approximate controller is derived, there is no straightforward way to trace the reasons for its proposed actions back to its inputs, hence making the controller a black-box model. In this paper, we propose the use of SHAP, an explainable artifical intelligence technique, to generate insights from learning-based MPC for the purpose of model debugging and simplification. Our results show that SHAP can explain general control behaviors and can also support model simplification in an informed way, representing a better alternative to dimensionality reduction techniques such as principal component analysis.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125302456","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931853
Natalie Pham, W. Gray
Given a smooth finite-dimensional state space system which is linear in the input, its input-output map can be represented by a Chen-Fliess series, namely, a weighted sum of iterated integrals of the input's component functions. The objective of this paper is to propose a generalized notion of a Chen- Fliess series for infinite-dimensional systems. The basic idea is to replace the real field of series coefficients with a ring of linear operators which act on the iterated integrals. The specific goals are to provide sufficient conditions for convergence of this generalized series and to exercise the theory on two specific examples: the transport equation and second-order hyperbolic partial differential equations. It will be shown in these examples that the generalized Chen-Fliess series under suitable conditions yields solutions that converge pointwise to the known classical solutions.
{"title":"Chen-Fliess Series for Linear Distributed Systems with One Spatial Dimension","authors":"Natalie Pham, W. Gray","doi":"10.1109/ICSTCC55426.2022.9931853","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931853","url":null,"abstract":"Given a smooth finite-dimensional state space system which is linear in the input, its input-output map can be represented by a Chen-Fliess series, namely, a weighted sum of iterated integrals of the input's component functions. The objective of this paper is to propose a generalized notion of a Chen- Fliess series for infinite-dimensional systems. The basic idea is to replace the real field of series coefficients with a ring of linear operators which act on the iterated integrals. The specific goals are to provide sufficient conditions for convergence of this generalized series and to exercise the theory on two specific examples: the transport equation and second-order hyperbolic partial differential equations. It will be shown in these examples that the generalized Chen-Fliess series under suitable conditions yields solutions that converge pointwise to the known classical solutions.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114347144","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931878
Ioana Hustiu, C. Mahulea, M. Kloetzer
This work considers the problem of decomposing a global motion specification given to a team of robots into parts that can be individually followed by agents, without any need of communication among them. The specification is given as a co-safe Linear Temporal Logic (LTL) formula over some regions of interest from a known and static environment. The main contribution is an algorithmic method for distributing the specification, while identifying necessary sequencing relations among some regions to be visited. Each of these sequences has to be assigned to a robot, and the concurrent movement of the agents accomplishes the global imposed mission.
{"title":"Distributing Co-safe LTL Specifications to Mobile Robots","authors":"Ioana Hustiu, C. Mahulea, M. Kloetzer","doi":"10.1109/ICSTCC55426.2022.9931878","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931878","url":null,"abstract":"This work considers the problem of decomposing a global motion specification given to a team of robots into parts that can be individually followed by agents, without any need of communication among them. The specification is given as a co-safe Linear Temporal Logic (LTL) formula over some regions of interest from a known and static environment. The main contribution is an algorithmic method for distributing the specification, while identifying necessary sequencing relations among some regions to be visited. Each of these sequences has to be assigned to a robot, and the concurrent movement of the agents accomplishes the global imposed mission.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114855299","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931862
Florin Sandru, Vlad-Ilie Ungureanu, I. Silea
The storage of vehicles, while they are not in operation, is one of the main challenges that infrastructure providers need to handle. The growth of personal vehicle ownership is directly impacting the demand for parking solutions, especially near areas of high interest. While an increase in capacity will require structural changes to the facilities themselves, a change in the mode the facility is used is easier to achieve. The way traffic is distributed among its area can decrease the downtime a parking resource is available and or the environmental footprint of its usage. The following paper presents a solution based on a mechanism traditionally used for distributing work among computing resources and its necessities for deployment. The presented solution provides minimal functionality even if data sources are unavailable thus allowing permanent service availability.
{"title":"Parking facility management solution based on dynamic traffic distribution","authors":"Florin Sandru, Vlad-Ilie Ungureanu, I. Silea","doi":"10.1109/ICSTCC55426.2022.9931862","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931862","url":null,"abstract":"The storage of vehicles, while they are not in operation, is one of the main challenges that infrastructure providers need to handle. The growth of personal vehicle ownership is directly impacting the demand for parking solutions, especially near areas of high interest. While an increase in capacity will require structural changes to the facilities themselves, a change in the mode the facility is used is easier to achieve. The way traffic is distributed among its area can decrease the downtime a parking resource is available and or the environmental footprint of its usage. The following paper presents a solution based on a mechanism traditionally used for distributing work among computing resources and its necessities for deployment. The presented solution provides minimal functionality even if data sources are unavailable thus allowing permanent service availability.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115008370","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 : 2022-10-19DOI: 10.1109/ICSTCC55426.2022.9931837
T. Ionescu, Lahcen El Bourkhissi, I. Necoara
In this paper, we study the problem of time-domain least squares moment matching-based model order reduction of linear systems. We first present the definition and the charac-terization of a model of order $r$ matching $rll nu$ moments of the given system. We then present the associated least squares moment matching problem in the form of a (nonconvex) optimization problem. Different from the existing results, we leave the interpolation points as decision variables and obtain an optimization problem with bilinear cost and constraints. The solution of the nonlinear least squares model reduction problem is computed at the optimal interpolation points using the efficient sequential convex programming algorithm. The proposed approach has practical advantages, since powerful convex optimization solvers, such as CVX, can be used to solve iteratively the optimization problem. A numerical example is given to illustrate the efficiency of our approach.
{"title":"Least squares moment matching-based model reduction using convex optimization","authors":"T. Ionescu, Lahcen El Bourkhissi, I. Necoara","doi":"10.1109/ICSTCC55426.2022.9931837","DOIUrl":"https://doi.org/10.1109/ICSTCC55426.2022.9931837","url":null,"abstract":"In this paper, we study the problem of time-domain least squares moment matching-based model order reduction of linear systems. We first present the definition and the charac-terization of a model of order $r$ matching $rll nu$ moments of the given system. We then present the associated least squares moment matching problem in the form of a (nonconvex) optimization problem. Different from the existing results, we leave the interpolation points as decision variables and obtain an optimization problem with bilinear cost and constraints. The solution of the nonlinear least squares model reduction problem is computed at the optimal interpolation points using the efficient sequential convex programming algorithm. The proposed approach has practical advantages, since powerful convex optimization solvers, such as CVX, can be used to solve iteratively the optimization problem. A numerical example is given to illustrate the efficiency of our approach.","PeriodicalId":220845,"journal":{"name":"2022 26th International Conference on System Theory, Control and Computing (ICSTCC)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133781170","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：