The Experts below are selected from a list of 285 Experts worldwide ranked by ideXlab platform
Frank L. Lewis - One of the best experts on this subject based on the ideXlab platform.
-
Coupled and constrained sylvester equations in system design
Circuits Systems and Signal Processing, 1994Co-Authors: Vassilis L. Syrmos, Frank L. LewisAbstract:Several design problems, including reduced observer and compensator design, output feedback, and finite transmission zero assignment, are examined using the vehicle of the coupled Sylvester equations. The coupling is generally provided through a third equation involving the solutions of the two linear Sylvester equations, thus serving as a constraint on the allowed solutions. The Sylvester approach allows the unification of algebraic and geometric approaches, and provides numerical design algorithms through the tool of the Hessenberg Form.
-
Transmission zero assignment using semistate descriptions
IEEE Transactions on Automatic Control, 1993Co-Authors: Vassilis L. Syrmos, Frank L. LewisAbstract:The problem of finite/infinite transmission zero assignment is examined by squaring a system from the outputs to the inputs. In particular, this problem is studied in two cases: state-accessible systems and partially-state accessible systems. It is shown that the problem of transmission zero assignment for state-accessible state-variable systems is equivalent to a pole-placement problem with state feedback for a generalized system, which always has a solution. In the case of partially-state-accessible systems, it is shown that the transmission zero assignment problem is equivalent to a pole-placement problem with output feedback for a generalized system. In both cases the block Hessenberg Form of the system and the extended lower triangular Hessenberg Form are exploited to Formulate the problem. >
-
Transmission Zero Assignment using Semistate Descriptions
1992 American Control Conference, 1992Co-Authors: Vassilis L. Syrmos, Frank L. LewisAbstract:In this paper the problem of finite/infinite transmission zero assignment is examined by squaring a system from the outputs to the inputs. In particular, we study this problem in two cases, state-accessible systems and partially state-accessible systems. We show that the problem of transmission zero assignment for state-accessible state-variable systems is equivalent to a pole-placement problem with state feedback for a generalized system, which always has a solution. In the case of partially state-accessible systems, we show that the transmission zero assignment problem is equivalent to a pole-placement problem with output feedback for a generalized system. In both cases we exploit the block Hessenberg Form of the system and the extended lower triangular Hessenberg Form in order to Formulate this problem.
-
State feedback design using reduced-order nonsquare descriptions
IEEE Transactions on Automatic Control, 1992Co-Authors: V.l. Syrmos, Frank L. LewisAbstract:A pole-placement technique is proposed for computing state feedback in multi-input linear systems. The algorithm has good numerical behavior and utilizes nonsquare pencils and the notion of decomposability. The fact that the method deals with reduced-order pencils and exploits the Schur-Hessenberg Form improves the conditioning of the problem. >
Vassilis L. Syrmos - One of the best experts on this subject based on the ideXlab platform.
-
External Descriptions and Staircase Forms in Implicit Systems
SIAM Journal on Matrix Analysis and Applications, 1995Co-Authors: Vassilis L. Syrmos, Petr Zagalak, Vladimír KučeraAbstract:This paper studies the relationship between staircase Forms and normal external descriptions in implicit systems. The authors show how to compute the proper and nonproper controllability indices of an implicit system using the reachability Hessenberg Form and the corresponding normal external description. The normal external description of the system is computed using embedding techniques. Finally, the differences and similarities between the normal external descriptions computed are shown using the reachability and controllability Hessenberg Forms of an implicit system.
-
Coupled and constrained sylvester equations in system design
Circuits Systems and Signal Processing, 1994Co-Authors: Vassilis L. Syrmos, Frank L. LewisAbstract:Several design problems, including reduced observer and compensator design, output feedback, and finite transmission zero assignment, are examined using the vehicle of the coupled Sylvester equations. The coupling is generally provided through a third equation involving the solutions of the two linear Sylvester equations, thus serving as a constraint on the allowed solutions. The Sylvester approach allows the unification of algebraic and geometric approaches, and provides numerical design algorithms through the tool of the Hessenberg Form.
-
Computing Normal External Descriptions and Feedback Design
1993Co-Authors: Vassilis L. Syrmos, Petr ZagalakAbstract:Given a linear system ?= Ax + Bu, we compute a normal external description (N(s), D(s)), using the Hessenberg Form of the pair (A, B) and embedding techniques. We show how to compute a state feedback K that assigns the closed-loop invariant polynomials using a Diophantine equation. The solution to such an equation corresponds to a back substitution problem due to the special structure of the computed normal external description. The proposed algorithms are easy to implement and computationally efficient.
-
Transmission zero assignment using semistate descriptions
IEEE Transactions on Automatic Control, 1993Co-Authors: Vassilis L. Syrmos, Frank L. LewisAbstract:The problem of finite/infinite transmission zero assignment is examined by squaring a system from the outputs to the inputs. In particular, this problem is studied in two cases: state-accessible systems and partially-state accessible systems. It is shown that the problem of transmission zero assignment for state-accessible state-variable systems is equivalent to a pole-placement problem with state feedback for a generalized system, which always has a solution. In the case of partially-state-accessible systems, it is shown that the transmission zero assignment problem is equivalent to a pole-placement problem with output feedback for a generalized system. In both cases the block Hessenberg Form of the system and the extended lower triangular Hessenberg Form are exploited to Formulate the problem. >
-
Transmission Zero Assignment using Semistate Descriptions
1992 American Control Conference, 1992Co-Authors: Vassilis L. Syrmos, Frank L. LewisAbstract:In this paper the problem of finite/infinite transmission zero assignment is examined by squaring a system from the outputs to the inputs. In particular, we study this problem in two cases, state-accessible systems and partially state-accessible systems. We show that the problem of transmission zero assignment for state-accessible state-variable systems is equivalent to a pole-placement problem with state feedback for a generalized system, which always has a solution. In the case of partially state-accessible systems, we show that the transmission zero assignment problem is equivalent to a pole-placement problem with output feedback for a generalized system. In both cases we exploit the block Hessenberg Form of the system and the extended lower triangular Hessenberg Form in order to Formulate this problem.
Bo Kågström - One of the best experts on this subject based on the ideXlab platform.
-
Evaluation of the Tunability of a New NUMA-Aware Hessenberg Reduction Algorithm
2016Co-Authors: Mahmoud Eljammaly, Lars Karlsson, Bo KågströmAbstract:The reduction of a general dense and square matrix to Hessenberg Form is a well known first step in many standard eigenvalue solvers. Although parallel algorithms exist, the Hessenberg reduction is ...
-
PARA (2) - Efficient reduction from block Hessenberg Form to Hessenberg Form using shared memory
Applied Parallel and Scientific Computing, 2012Co-Authors: Lars Karlsson, Bo KågströmAbstract:A new cache-efficient algorithm for reduction from block Hessenberg Form to Hessenberg Form is presented and evaluated. The algorithm targets parallel computers with shared memory. One level of look-ahead in combination with a dynamic load-balancing scheme significantly reduces the idle time and allows the use of coarse-grained tasks. The coarse tasks lead to high-perFormance computations on each processor/core. Speedups close to 13 over the sequential unblocked algorithm have been observed on a dual quad-core machine using one thread per core.
-
parallel two stage reduction to Hessenberg Form using dynamic scheduling on shared memory architectures
Parallel Computing, 2011Co-Authors: Lars Karlsson, Bo KågströmAbstract:We consider parallel reduction of a real matrix to Hessenberg Form using orthogonal transFormations. Standard Hessenberg reduction algorithms reduce the columns of the matrix from left to right in either a blocked or unblocked fashion. However, the standard blocked variant perForms 20% of the computations in terms of matrix-vector multiplications. We show that a two-stage approach consisting of an intermediate reduction to block Hessenberg Form speeds up the reduction by avoiding matrix-vector multiplications. We describe and evaluate a new high-perFormance implementation of the two-stage approach that attains significant speedups over the one-stage approach. The key components are a dynamically scheduled implementation of Stage 1 and a blocked, adaptively load-balanced implementation of Stage 2.
-
efficient reduction from block Hessenberg Form to Hessenberg Form using shared memory
Parallel Computing, 2010Co-Authors: Lars Karlsson, Bo KågströmAbstract:A new cache-efficient algorithm for reduction from block Hessenberg Form to Hessenberg Form is presented and evaluated. The algorithm targets parallel computers with shared memory. One level of look-ahead in combination with a dynamic load-balancing scheme significantly reduces the idle time and allows the use of coarse-grained tasks. The coarse tasks lead to high-perFormance computations on each processor/core. Speedups close to 13 over the sequential unblocked algorithm have been observed on a dual quad-core machine using one thread per core.
-
Parallel two-stage reduction to Hessenberg Form using shared memory
2010Co-Authors: Lars Karlsson, Bo KågströmAbstract:We consider parallel reduction of a real matrix to Hessenberg Form using orthogonal transFormations. Standard Hessenberg reduction algorithms reduce the columns of the matrix from left to right in either a blocked or unblocked fashion. However, the standard blocked variant perForms 20% of the computations as large matrix‐vector multiplications. We show that a two-stage approach consisting of an intermediate reduction to r-Hessenberg Form speeds up the reduction by avoiding matrix‐vector multiplications. We describe and evaluate a new high-perFormance implementation of the two-stage approach that attains significant speedups over the one-stage approach on a dual quad-core machine. The key components are a dynamically scheduled implementation of the first stage and a blocked adaptively load-balanced implementation of the second stage.
V.l. Syrmos - One of the best experts on this subject based on the ideXlab platform.
-
Computational design techniques for reduced-order observers in generalized state-space systems
Circuits Systems and Signal Processing, 1993Co-Authors: V.l. SyrmosAbstract:In this paper we present computationally efficient design techniques for the reduced-order observer problem in generalized state-space systems. The proposed approach utilizes the generalized block Hessenberg Form. The problem of reduced-order observer in generalized systems is reFormulated as a full-order observer problem for statespace systems. This new approach parametrizes the family of all possible solutions, studies the problem of robustness, and relaxes the condition of observability.
-
State feedback design using reduced-order nonsquare descriptions
IEEE Transactions on Automatic Control, 1992Co-Authors: V.l. Syrmos, Frank L. LewisAbstract:A pole-placement technique is proposed for computing state feedback in multi-input linear systems. The algorithm has good numerical behavior and utilizes nonsquare pencils and the notion of decomposability. The fact that the method deals with reduced-order pencils and exploits the Schur-Hessenberg Form improves the conditioning of the problem. >
Lars Karlsson - One of the best experts on this subject based on the ideXlab platform.
-
Evaluation of the Tunability of a New NUMA-Aware Hessenberg Reduction Algorithm
2016Co-Authors: Mahmoud Eljammaly, Lars Karlsson, Bo KågströmAbstract:The reduction of a general dense and square matrix to Hessenberg Form is a well known first step in many standard eigenvalue solvers. Although parallel algorithms exist, the Hessenberg reduction is ...
-
PARA (2) - Efficient reduction from block Hessenberg Form to Hessenberg Form using shared memory
Applied Parallel and Scientific Computing, 2012Co-Authors: Lars Karlsson, Bo KågströmAbstract:A new cache-efficient algorithm for reduction from block Hessenberg Form to Hessenberg Form is presented and evaluated. The algorithm targets parallel computers with shared memory. One level of look-ahead in combination with a dynamic load-balancing scheme significantly reduces the idle time and allows the use of coarse-grained tasks. The coarse tasks lead to high-perFormance computations on each processor/core. Speedups close to 13 over the sequential unblocked algorithm have been observed on a dual quad-core machine using one thread per core.
-
parallel two stage reduction to Hessenberg Form using dynamic scheduling on shared memory architectures
Parallel Computing, 2011Co-Authors: Lars Karlsson, Bo KågströmAbstract:We consider parallel reduction of a real matrix to Hessenberg Form using orthogonal transFormations. Standard Hessenberg reduction algorithms reduce the columns of the matrix from left to right in either a blocked or unblocked fashion. However, the standard blocked variant perForms 20% of the computations in terms of matrix-vector multiplications. We show that a two-stage approach consisting of an intermediate reduction to block Hessenberg Form speeds up the reduction by avoiding matrix-vector multiplications. We describe and evaluate a new high-perFormance implementation of the two-stage approach that attains significant speedups over the one-stage approach. The key components are a dynamically scheduled implementation of Stage 1 and a blocked, adaptively load-balanced implementation of Stage 2.
-
efficient reduction from block Hessenberg Form to Hessenberg Form using shared memory
Parallel Computing, 2010Co-Authors: Lars Karlsson, Bo KågströmAbstract:A new cache-efficient algorithm for reduction from block Hessenberg Form to Hessenberg Form is presented and evaluated. The algorithm targets parallel computers with shared memory. One level of look-ahead in combination with a dynamic load-balancing scheme significantly reduces the idle time and allows the use of coarse-grained tasks. The coarse tasks lead to high-perFormance computations on each processor/core. Speedups close to 13 over the sequential unblocked algorithm have been observed on a dual quad-core machine using one thread per core.
-
Parallel two-stage reduction to Hessenberg Form using shared memory
2010Co-Authors: Lars Karlsson, Bo KågströmAbstract:We consider parallel reduction of a real matrix to Hessenberg Form using orthogonal transFormations. Standard Hessenberg reduction algorithms reduce the columns of the matrix from left to right in either a blocked or unblocked fashion. However, the standard blocked variant perForms 20% of the computations as large matrix‐vector multiplications. We show that a two-stage approach consisting of an intermediate reduction to r-Hessenberg Form speeds up the reduction by avoiding matrix‐vector multiplications. We describe and evaluate a new high-perFormance implementation of the two-stage approach that attains significant speedups over the one-stage approach on a dual quad-core machine. The key components are a dynamically scheduled implementation of the first stage and a blocked adaptively load-balanced implementation of the second stage.