Madhu N. Belur
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Madhu N. Belur has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Prasanna Chaporkar
Department of Electrical Engineering Indian Institute of Technology Bombay India
6 shared publications - Shana Moothedath
Department of Electrical and Computer Engineering, Iowa State University, Ames, IA, USA
5 shared publications - Ashish Kothyari
Indian Institute of Technology Roorkee
3 shared publications Debasattam Pal
Department of Electrical Engineering, Indian Institute of Technology Bombay, India
3 shared publications- Rachel Kalpana Kalaimani
Department of Electrical Engineering, Indian Institute of Technology Madras, Chennai, India
3 shared publications - Chayan Bhawal
Department of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai, India
2 shared publications
Latest Publications
- 2 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 10 (3) , 1636-1647
Algebraic Connectivity: Local and Global Maximizer Graphs
Algebraic connectivity is one way to quantify graph connectivity, which in turn gauges robustness as a network. In this paper, we consider the problem of maximizing algebraic connectivity both locally and globally overall simple, undirected, unweighted graphs with a given number of vertices and edges. We pursue this optimization by equivalently minimizing the largest eigenvalue of the Laplacian of the ‘complement graph’. We establish that the union of complete subgraphs are largest eigenvalue local minimizer graphs. Further, under sufficient conditions satisfied by the edge/vertex counts, we prove that this union of complete components graphs are, in fact, Laplacian largest eigenvalue global maximizers; these results generalize the ones in the literature that are for just two components. These sufficient conditions can be viewed as quantifying situations where the component sizes are either ‘quite homogeneous’ or some of them are relatively ‘negligibly small,’ and thus generalize known results of homogeneity of components. While a conjecture about global optimality of complete bipartite graphs' from the literature continues to remain open, assuming appropriate constraints we prove the conjecture and also formulate/prove a variant of this claim. We finally relate this central optimization problem in this paper with the Discrete Fourier Transform (DFT) and circulant graphs/matrices.
- 2 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 10 (3) , 1636-1647
Algebraic Connectivity: Local and Global Maximizer Graphs
Algebraic connectivity is one way to quantify graph connectivity, which in turn gauges robustness as a network. In this paper, we consider the problem of maximizing algebraic connectivity both locally and globally overall simple, undirected, unweighted graphs with a given number of vertices and edges. We pursue this optimization by equivalently minimizing the largest eigenvalue of the Laplacian of the ‘complement graph’. We establish that the union of complete subgraphs are largest eigenvalue local minimizer graphs. Further, under sufficient conditions satisfied by the edge/vertex counts, we prove that this union of complete components graphs are, in fact, Laplacian largest eigenvalue global maximizers; these results generalize the ones in the literature that are for just two components. These sufficient conditions can be viewed as quantifying situations where the component sizes are either ‘quite homogeneous’ or some of them are relatively ‘negligibly small,’ and thus generalize known results of homogeneity of components. While a conjecture about global optimality of complete bipartite graphs' from the literature continues to remain open, assuming appropriate constraints we prove the conjecture and also formulate/prove a variant of this claim. We finally relate this central optimization problem in this paper with the Discrete Fourier Transform (DFT) and circulant graphs/matrices.
- 5 October 2022
- Published by Taylor & Francis in International Journal of Control
- Vol. 96 (12) , 2958-2975
Imaginary axis eigenvalues of Hamiltonian matrix: controllability, defectiveness and the ϵ-characteristic
The eigenstructure of imaginary axis eigenvalues of a Hamiltonian matrix is of importance in many fields of control systems, for example, in stability analysis of linear Hamiltonian systems, computation of the solutions of an algebraic Riccati equation (ARE), existence of a Lyapunov function for LTI systems, etc. The dynamical system consisting of all stationary trajectories for an optimal control problem – often called the Hamiltonian system – is known to admit, in its state space dynamical equation, a Hamiltonian matrix for the system matrix. In each of these cases, defectiveness of imaginary axis eigenvalues of a Hamiltonian matrix turns out to be of crucial importance. For example, defectiveness causes unboundedness of the oscillatory stationary trajectories in the Hamiltonian system. A characterisation of the ARE solutions in terms of Lagrangian invariant subspaces of the Hamiltonian matrix when the imaginary axis eigenvalues are defective has been addressed in the literature for the controllable case. This paper focuses on the general case of uncontrollable systems; we formulate conditions under which the imaginary axis eigenvalues of the Hamiltonian matrix are non-defective: this is central for solutions corresponding to imaginary axis eigenvalues to not become unbounded. We provide conditions on the so-called -characteristic of the non-defective imaginary axis eigenvalue: this helps in the characterisation of Lagrangian invariant subspaces. We formulate conditions under which a passivity-based Hamiltonian matrix is normal (i.e. it commutes with its transpose), and we link normality of such Hamiltonian matrices with all-pass behaviour. In summary, this paper formulates results that link defectiveness of imaginary axis eigenvalues of Hamiltonian matrix to solvabilities of Lyapunov and Algebraic Riccati equations, controllability/observability, -characteristic and sign-controllability. We consider examples in the area of bounded-real transfer functions and RLC circuits, both controllable and uncontrollable, to study applicability of the results of this paper.- 5 October 2022
- Published by Taylor & Francis in International Journal of Control
- Vol. 96 (12) , 2958-2975
Imaginary axis eigenvalues of Hamiltonian matrix: controllability, defectiveness and the ϵ-characteristic
The eigenstructure of imaginary axis eigenvalues of a Hamiltonian matrix is of importance in many fields of control systems, for example, in stability analysis of linear Hamiltonian systems, computation of the solutions of an algebraic Riccati equation (ARE), existence of a Lyapunov function for LTI systems, etc. The dynamical system consisting of all stationary trajectories for an optimal control problem – often called the Hamiltonian system – is known to admit, in its state space dynamical equation, a Hamiltonian matrix for the system matrix. In each of these cases, defectiveness of imaginary axis eigenvalues of a Hamiltonian matrix turns out to be of crucial importance. For example, defectiveness causes unboundedness of the oscillatory stationary trajectories in the Hamiltonian system. A characterisation of the ARE solutions in terms of Lagrangian invariant subspaces of the Hamiltonian matrix when the imaginary axis eigenvalues are defective has been addressed in the literature for the controllable case. This paper focuses on the general case of uncontrollable systems; we formulate conditions under which the imaginary axis eigenvalues of the Hamiltonian matrix are non-defective: this is central for solutions corresponding to imaginary axis eigenvalues to not become unbounded. We provide conditions on the so-called -characteristic of the non-defective imaginary axis eigenvalue: this helps in the characterisation of Lagrangian invariant subspaces. We formulate conditions under which a passivity-based Hamiltonian matrix is normal (i.e. it commutes with its transpose), and we link normality of such Hamiltonian matrices with all-pass behaviour. In summary, this paper formulates results that link defectiveness of imaginary axis eigenvalues of Hamiltonian matrix to solvabilities of Lyapunov and Algebraic Riccati equations, controllability/observability, -characteristic and sign-controllability. We consider examples in the area of bounded-real transfer functions and RLC circuits, both controllable and uncontrollable, to study applicability of the results of this paper.- 24 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 67 (7) , 3746-3753
Interlacing Properties of System-Poles, System-Zeros, and Spectral-Zeros in MIMO Systems
SISO systems with zeros interlacing poles (ZIP) have been extensively studied in the literature and have received interest widely. However, the ZIP property for MIMO systems has not been pursued sufficiently. In this paper we pursue the ZIP property for MIMO systems, and link this to strict passivity. For the class of systems that admit a symmetric state-space realization, which for the SISO case is equivalent to ZIP, we obtain sufficient conditions under which MIMO systems too have ZIP. We also present new results in the context of ‘spectral-zero’ of a system, a notion that plays a key role in many optimal control and estimation problems. We formulate conditions under which the spectral-zeros of a MIMO system are real, and further, conditions that guarantee that the system-zeros, spectral-zeros and the system-poles are all interlaced for MIMO systems.
- 24 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 67 (7) , 3746-3753
Interlacing Properties of System-Poles, System-Zeros, and Spectral-Zeros in MIMO Systems
SISO systems with zeros interlacing poles (ZIP) have been extensively studied in the literature and have received interest widely. However, the ZIP property for MIMO systems has not been pursued sufficiently. In this paper we pursue the ZIP property for MIMO systems, and link this to strict passivity. For the class of systems that admit a symmetric state-space realization, which for the SISO case is equivalent to ZIP, we obtain sufficient conditions under which MIMO systems too have ZIP. We also present new results in the context of ‘spectral-zero’ of a system, a notion that plays a key role in many optimal control and estimation problems. We formulate conditions under which the spectral-zeros of a MIMO system are real, and further, conditions that guarantee that the system-zeros, spectral-zeros and the system-poles are all interlaced for MIMO systems.
- 14 January 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Control of Network Systems
- Vol. 7 (3) , 1164-1175
Optimal Network Topology Design in Composite Systems for Structural Controllability
- 14 January 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Control of Network Systems
- Vol. 7 (3) , 1164-1175
Optimal Network Topology Design in Composite Systems for Structural Controllability
- 4 June 2019
- Published by Elsevier in Automatica
- Vol. 107, 200-210
Approximating constrained minimum cost input–output selection for generic arbitrary pole placement in structured systems
- 4 June 2019
- Published by Elsevier in Automatica
- Vol. 107, 200-210
Approximating constrained minimum cost input–output selection for generic arbitrary pole placement in structured systems
- 1 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 260-265
Development of a railway junction simulator for evaluation of control strategies and capacity utilization optimization
This paper focuses on the development of a Python based tool for analysis and simulation of mixed rail traffic at a railway junction. The tool is helpful for evaluating various open-loop and feedback based strategies for an optimum junction utilization. The tool is implemented in Python using graph theory for finding shortest paths between various source-destination pairs. What signifies a railway junction, as opposed to a railway section is the possibility and the need to allow multiple simultaneous movements within the junction area. The tool applies network (edge and node) based approach for designing the line infrastructure of a railway junction. Source and destination are represented as nodes and every path from a source to a destination through a junction consists of a set of permissible combination of edges and nodes. This tool can be used for simulating scheduled passenger trains through all the allowed paths, and then selecting the best path for traversing through the junction. Further, lower priority trains are scheduled without affecting the edge/node occupancies trains of higher priorities. The best freight paths are identified and a more accurate estimate of the time required for freight trains to pass through the junction is obtained using this tool. In this way, one is able to quantify and then optimize the capacity/utilization of railway junction. In a larger scenario, the tool is useful for identifying bottlenecks in a given infrastructure. As examples, a detailed analysis of Allahabad junction and Kanpur junction of Indian Railways is presented using the tool.
- 1 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 260-265
Development of a railway junction simulator for evaluation of control strategies and capacity utilization optimization
This paper focuses on the development of a Python based tool for analysis and simulation of mixed rail traffic at a railway junction. The tool is helpful for evaluating various open-loop and feedback based strategies for an optimum junction utilization. The tool is implemented in Python using graph theory for finding shortest paths between various source-destination pairs. What signifies a railway junction, as opposed to a railway section is the possibility and the need to allow multiple simultaneous movements within the junction area. The tool applies network (edge and node) based approach for designing the line infrastructure of a railway junction. Source and destination are represented as nodes and every path from a source to a destination through a junction consists of a set of permissible combination of edges and nodes. This tool can be used for simulating scheduled passenger trains through all the allowed paths, and then selecting the best path for traversing through the junction. Further, lower priority trains are scheduled without affecting the edge/node occupancies trains of higher priorities. The best freight paths are identified and a more accurate estimate of the time required for freight trains to pass through the junction is obtained using this tool. In this way, one is able to quantify and then optimize the capacity/utilization of railway junction. In a larger scenario, the tool is useful for identifying bottlenecks in a given infrastructure. As examples, a detailed analysis of Allahabad junction and Kanpur junction of Indian Railways is presented using the tool.
- 26 July 2018
- Published by Taylor & Francis in International Journal of Control
- Vol. 93 (5) , 1204-1217
Closed-form solutions of singular KYP lemma: strongly passive systems, and fast lossless trajectories
In this paper, we deal with a special class of passive systems, which possess the characteristic property of having no finite spectral zeros. We call these systems strongly passive. It is well known that, for these systems, storage functions, i.e. solutions to the linear matrix inequality (LMI) arising from the Kalman–Yakubovich–Popov (KYP) lemma, cannot be obtained by the conventional approach of algebraic Riccati equations (AREs) and Hamiltonian matrices. In this paper, we first show that a strongly passive system always admits a unique storage function. We then provide a closed-form expression for this unique storage function. Using the closed-form formula of the unique storage function we characterise the ‘lossless’ trajectories of strongly passive systems and show that such systems admit impulsive lossless trajectories on the half-line; we call them fast lossless trajectories. This adds to the existing notion that such systems do not admit any ‘slow’ lossless trajectories.
- 26 July 2018
- Published by Taylor & Francis in International Journal of Control
- Vol. 93 (5) , 1204-1217
Closed-form solutions of singular KYP lemma: strongly passive systems, and fast lossless trajectories
In this paper, we deal with a special class of passive systems, which possess the characteristic property of having no finite spectral zeros. We call these systems strongly passive. It is well known that, for these systems, storage functions, i.e. solutions to the linear matrix inequality (LMI) arising from the Kalman–Yakubovich–Popov (KYP) lemma, cannot be obtained by the conventional approach of algebraic Riccati equations (AREs) and Hamiltonian matrices. In this paper, we first show that a strongly passive system always admits a unique storage function. We then provide a closed-form expression for this unique storage function. Using the closed-form formula of the unique storage function we characterise the ‘lossless’ trajectories of strongly passive systems and show that such systems admit impulsive lossless trajectories on the half-line; we call them fast lossless trajectories. This adds to the existing notion that such systems do not admit any ‘slow’ lossless trajectories.
- 30 May 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 65 (12) , 4349-4362
Lossless Systems Storage Function: New Results and Numerically Stable and Non-Iterative Computational Methods
In this paper, we formulate and prove new results in the context of storage functions for lossless systems: we use these results to propose new algorithms to compute the storage function. The computation of the storage function for the lossless case is not possible using conventional algebraic Riccati equation-based algorithms, though the storage function itself is well-defined. This is because a certain ``regularity condition'' on the feedthrough term in the i/s/o representation of the lossless system does not hold. We formulate new results about the storage function matrix for the lossless case and use them to propose non-iterative and stable algorithms to compute the storage function directly from different representations of the given system, namely, a kernel representation, transfer function, and the i/s/o representation of the system. Across the methods, for randomly generated transfer functions, we compare: 1) the computational effort (in flops); 2) the computation time using numerical experiments; and 3) the computational error.
- 30 May 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 65 (12) , 4349-4362
Lossless Systems Storage Function: New Results and Numerically Stable and Non-Iterative Computational Methods
In this paper, we formulate and prove new results in the context of storage functions for lossless systems: we use these results to propose new algorithms to compute the storage function. The computation of the storage function for the lossless case is not possible using conventional algebraic Riccati equation-based algorithms, though the storage function itself is well-defined. This is because a certain ``regularity condition'' on the feedthrough term in the i/s/o representation of the lossless system does not hold. We formulate new results about the storage function matrix for the lossless case and use them to propose non-iterative and stable algorithms to compute the storage function directly from different representations of the given system, namely, a kernel representation, transfer function, and the i/s/o representation of the system. Across the methods, for randomly generated transfer functions, we compare: 1) the computational effort (in flops); 2) the computation time using numerical experiments; and 3) the computational error.
- 27 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 64 (3) , 1137-1149
On the Distance to Singular Descriptor Dynamical Systems With Impulsive Initial Conditions
This paper studies the problem of computing the distance between a given singular descriptor system $(E,A)$, and a nearest descriptor system that has impulsive initial conditions. Much of the literature focusses on the case when only one of $E$ and $A$ is perturbed. We give a closed form expression of the distance to a nearest descriptor system having impulsive solutions via rank-1 perturbations when both $E$ and $A$ are perturbed. Next, for the case of perturbations without rank restrictions, we propose and evaluate the bounds for the distance. In the context of structured perturbations, we formulate and obtain an explicit expression for the distance, when $E$ and $A$ are Hermitian and are perturbed by Hermitian matrices. For a suitable class of systems, we also show that the upper/lower bounds are within a factor of $\sqrt{2}$. We also construct examples and compare our bounds with those from the literature as well as with computed values of the distance obtained via three numerical optimization techniques like the Structured Low Rank Approximation (SLRA), the Broyden Fletcher Goldfarb Shanno (BFGS) algorithm, and direct optimization tools like globalsearch.- 27 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 64 (3) , 1137-1149
On the Distance to Singular Descriptor Dynamical Systems With Impulsive Initial Conditions
This paper studies the problem of computing the distance between a given singular descriptor system $(E,A)$, and a nearest descriptor system that has impulsive initial conditions. Much of the literature focusses on the case when only one of $E$ and $A$ is perturbed. We give a closed form expression of the distance to a nearest descriptor system having impulsive solutions via rank-1 perturbations when both $E$ and $A$ are perturbed. Next, for the case of perturbations without rank restrictions, we propose and evaluate the bounds for the distance. In the context of structured perturbations, we formulate and obtain an explicit expression for the distance, when $E$ and $A$ are Hermitian and are perturbed by Hermitian matrices. For a suitable class of systems, we also show that the upper/lower bounds are within a factor of $\sqrt{2}$. We also construct examples and compare our bounds with those from the literature as well as with computed values of the distance obtained via three numerical optimization techniques like the Structured Low Rank Approximation (SLRA), the Broyden Fletcher Goldfarb Shanno (BFGS) algorithm, and direct optimization tools like globalsearch.- 5 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 63 (11) , 3881-3888
Minimum Cost Feedback Selection for Arbitrary Pole Placement in Structured Systems
This paper addresses optimal feedback selection for arbitrary pole placement of structured systems when each feedback edge is associated with a cost. Given a structured system and a feedback cost matrix, our aim is to find a feasible feedback matrix of minimum cost that guarantees arbitrary pole placement of the closed-loop structured system. The same problem but with uniform costs has been considered recently, but the claim and the NP-hardness proof there has subtle flaws: we mention this briefly in our paper and then prove the NP-hardness of the feedback selection problem using a reduction from the weighted set cover problem. We next prove the polynomial time constant factor inapproximability of the problem by showing that a constant factor approximation for the problem does not exist, unless the weighted set cover problem can also be approximated within a constant factor. We study a subclass of systems whose directed acyclic graph constructed using the strongly connected components of the state digraph is a line graph and the state bipartite graph has a perfect matching. We propose a polynomial time algorithm based on dynamic programming principles for optimal feedba
- 5 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automatic Control
- Vol. 63 (11) , 3881-3888
Minimum Cost Feedback Selection for Arbitrary Pole Placement in Structured Systems
This paper addresses optimal feedback selection for arbitrary pole placement of structured systems when each feedback edge is associated with a cost. Given a structured system and a feedback cost matrix, our aim is to find a feasible feedback matrix of minimum cost that guarantees arbitrary pole placement of the closed-loop structured system. The same problem but with uniform costs has been considered recently, but the claim and the NP-hardness proof there has subtle flaws: we mention this briefly in our paper and then prove the NP-hardness of the feedback selection problem using a reduction from the weighted set cover problem. We next prove the polynomial time constant factor inapproximability of the problem by showing that a constant factor approximation for the problem does not exist, unless the weighted set cover problem can also be approximated within a constant factor. We study a subclass of systems whose directed acyclic graph constructed using the strongly connected components of the state digraph is a line graph and the state bipartite graph has a perfect matching. We propose a polynomial time algorithm based on dynamic programming principles for optimal feedba