Differential Algebraic Methods Research Articles

Polynomial Propagation of Moments in Stochastic Differential Equations

We address the problem of approximating the moments of the solution, , of an Itô stochastic differential equation (SDE) with drift and diffusion terms over a time grid . In particular, we assume an explicit numerical scheme for the generation of sample paths and then obtain recursive equations that yield any desired noncentral moment of as a function of the initial condition . The core of the methodology is the decomposition of the numerical solution into a “central part” and an “effective noise” term. The central term is computed deterministically from the ordinary differential equation (ODE) that results from eliminating the diffusion term in the SDE, while the effective noise accounts for the stochastic deviation from the numerical solution of the ODE. For simplicity, we describe the proposed methodology based on an Euler–Maruyama integrator, but other explicit numerical schemes can be exploited in the same way. We also apply the moment approximations to construct estimates of the 1-dimensional marginal probability density functions of based on a Gram–Charlier expansion. Both for the approximation of moments and 1-dimensional densities, we describe how to handle the cases in which the initial condition is fixed (i.e., for some deterministic and known ) or random. In the latter case, we resort to polynomial chaos expansion (PCE) schemes in order to approximate the target moments. The methodology has been inspired by the PCE and differential algebra methods used for uncertainty propagation in astrodynamics problems. Hence, we illustrate its application for the quantification of uncertainty in a 2-dimensional Keplerian orbit perturbed by a Wiener noise process.

TEMGYM Advanced: Software for electron lens aberrations and parallelised electron ray tracing

Characterisation of the electron beams trajectory in an electron microscope is possible in a few select commercial software packages, but these tools and their source code are not available in a free and accessible manner. This paper introduces the free and open-source software TEMGYM Advanced, which implements ray tracing methods that calculate the path of electrons through a magnetic or electrostatic lens and allow evaluation of the first-order properties and third-order geometric aberrations. Validation of the aberration coefficient calculations is performed by implementing two independent methods – the aberration integral and differential algebra (DA) methods and by comparing the results of each. This paper also demonstrates parallelised electron ray tracing through a series of magnetic components, which enables near real-time generation of a physically accurate beam-spot including aberrations and brings closer the realisation of a digital twin of an electron microscope. TEMGYM Advanced represents a valuable resource for the electron microscopy community, providing an accessible and open source means of characterising electron lenses. This software utilises the Python programming language to complement the growing ecosystem of free and open-source software within the electron microscopy community, and to facilitate the application of machine learning to an electron microscope digital twin for instrument automation. The software is available under GNU Public License number Three (GPL 3).

Differential algebra methods applied to continuous abacus generation and bifurcation detection: application to periodic families of the Earth–Moon system

The return of human space missions to the Moon puts the Earth–Moon system (EMS) at the center of attention. Hence, studying the periodic solutions to the circular restricted three-body problem (CR3BP) is crucial to ease transfer computations, find new solutions, or to better understand these orbits. This work proposes a novel continuation method of periodic families using differential algebra (DA) mapping. We exploit DA with automatic control of the truncation error to represent each family of periodic orbits as a set 2D Taylor polynomial maps. These maps guarantee the access to any point of the family without any numerical propagation, providing a continuous abacus. When applied to the halo family at $$L_1$$ and $$L_2$$ , the planar Lyapunov at $$L_1$$ and $$L_2$$ , the distant retrograde orbit (DRO) family, and the butterfly family, we show that the DA-based 2D mapping is asymptotically more efficient than point-wise methods by at least two orders of magnitude, with controlled accuracy. To assist the computation of family of periodic orbits, we propose a novel DA-based automatic bifurcation detection algorithm that enables the continuous mapping of the family’s bifurcation criteria. A bifurcation study on the halo $$L_2$$ shows identical results as point-wise methods while highlighting two undocumented bifurcations.

Differential algebraic fast multipole-accelerated boundary element method for nonlinear beam dynamics in arbitrary enclosures

A novel method is developed to take into account realistic boundary conditions in intense nonlinear beam dynamics. The algorithm consists of three main ingredients: the boundary element method that provides a solution for the discretized reformulation of the Poisson equation as boundary integrals; a novel fast multipole method developed for accurate and efficient computation of Coulomb potentials and forces; and differential algebraic methods, which form the numerical structures that enable and hold together the different components. The fast multipole method, without any modifications, also accelerates the solution of intertwining linear systems of equations for further efficiency enhancements. The resulting algorithm scales linearly with the number of particles $N$, as $m\text{ }\mathrm{log}\text{ }m$ with the number of boundary elements $m$, and, therefore, establishes an accurate and efficient method for intense beam dynamics simulations in arbitrary enclosures. Its performance is illustrated with three different cases and structures of practical interest.

COMBOS2: an algorithm to the input–output equations of dynamic biosystems via Gaussian elimination

Differential algebra (DA) methods are currently being exploited for analyzing dynamic biosystem models for their structural identifiability (SI) properties. An early step in this approach entails finding an equivalent input–output (I/O) model. A recent approach for finding these equations, based on Grbner bases and imbedded in the app COMBOS is relatively slow and sometimes unsuccessful, even for moderate size models. In this paper, we propose a faster algorithm, using a Gaussian elimination approach for analyzing the subclass of linear dynamic biosystem models. We apply this methodology to find the simplest set of globally SI parameter combinations of these models and show that it works effectively for linear biosystem models of simple to moderate complexity.

High order aberrations calculations of Wien filters using differential algebra methods

In this paper, a novel scheme based on the differential algebraic (DA) method is proposed to analyze the electron optics properties of the Wien filters. A new software package is then developed, to compute the geometrical and chromatic aberrations up to the fifth order of the Wien filters. For examples, the aberrations of single filter and the double filters are calculated. The calculated geometrical aberration coefficients are comparable with the counterpart calculated by Tang's theory. However, both theories can only match with each other in the calculation of the chromatic aberration coefficient when the dispersion ray is taken into consideration in Tang's theory.

Redundancy Relations for the Diagnosis of Hybrid Systems

In the concept of analytical redundancy, the problem of the functional diagnosis of hybrid systems is studied. The solution of this problem assumes checking the redundancy relations between the inputs and outputs of the system. A method for constructing the redundancy relations that includes transformation of the operational and controlling subsystems and then reduction of the operational subsystem to a form without feedbacks is proposed. The problem is solved using the pair algebra of partitions, algebra of functions, and differential-algebraic methods.

Algebraic estimation and active disturbance rejection in the control of flat systems

This paper proposes a feedback method for the control of uncertain systems with unknown external disturbances, which includes an algebraic estimator and relies on the Active Disturbance Rejection Control (ADRC) approach. The proposed estimator considers a generalized disturbance in order to deal with systems which may simultaneously present time varying parameters, external disturbances, un-modeled dynamics, and process noise. The on-line estimated disturbance is obtained by means of differential algebraic methods and it is used as the major part of an on-line feedback cancellation scheme aiming at linearization and uncertainty suppression. The algebraic estimator proposed in the paper makes unnecessary the use of classical extended state observers, which are widely used in ADRC. The speed of response and reliability of the proposed algebraic disturbance estimator-based control scheme was experimentally tested on three laboratory systems, including a system of directly-coupled DC motors, a roto-magnet system, and a disc and beam system, showing that the experimental results are in excellent agreement with the predictions of the theory.

Identifiability of mathematical models in medical biology

The analysis of biological data is a key topic in bioinformatics, computational genomics, molecular modeling, and systems biology. The methods covered in this article can reduce the cost of experiments aimed at obtaining biological data. The problem of the identifiability of mathematical models in physiology, pharmacokinetics, and epidemiology is considered. The processes considered are modeled using nonlinear systems of ordinary differential equations. The mathematical modeling of dynamic processes is based on the use of the mass conservation law. The problem of estimating the parameters characterizing the process under study raises the question of nonuniqueness. When the input and output data are known, it is useful to perform an a priori analysis of the relevance of these data. The definition of the identifiability of mathematical models is considered. Methods for the analysis of the identifiability of dynamic models are reviewed. In this review article, the following approaches are considered: the transfer function method applied to linear models (useful for the analysis of pharmacokinetic data, since a large class of drugs is characterized by linear kinetics); the Taylor series expansion method applied to nonlinear models; the differential algebra methods (the structure of this algorithm allows it to be run on a computer); and a method based on graph theory (this method allows for the analysis of the identifiability of the model and finding a proper reparametrization reducing the initial model to an identifiable one). The need to perform a priori identifiability analysis before estimating the parameters characterizing any process is demonstrated with several examples. The examples of identifiability analysis of mathematical models in medical biology are presented.

A Systematic Approach to Determining the Identifiability of Multistage Carcinogenesis Models.

Multistage clonal expansion (MSCE) models of carcinogenesis are continuous-time Markov process models often used to relate cancer incidence to biological mechanism. Identifiability analysis determines what model parameter combinations can, theoretically, be estimated from given data. We use a systematic approach, based on differential algebra methods traditionally used for deterministic ordinary differential equation (ODE) models, to determine identifiable combinations for a generalized subclass of MSCE models with any number of preinitation stages and one clonal expansion. Additionally, we determine the identifiable combinations of the generalized MSCE model with up to four clonal expansion stages, and conjecture the results for any number of clonal expansion stages. The results improve upon previous work in a number of ways and provide a framework to find the identifiable combinations for further variations on the MSCE models. Finally, our approach, which takes advantage of the Kolmogorov backward equations for the probability generating functions of the Markov process, demonstrates that identifiability methods used in engineering and mathematics for systems of ODEs can be applied to continuous-time Markov processes.

A Differential Algebraic Method for the Solution of the Poisson Equation for Charged Particle Beams

AbstractThe design optimization and analysis of charged particle beam systems employing intense beams requires a robust and accurate Poisson solver. This paper presents a new type of Poisson solver which allows the effects of space charge to be elegantly included into the system dynamics. This is done by casting the charge distribution function into a series of basis functions, which are then integrated with an appropriate Green's function to find a Taylor series of the potential at a given point within the desired distribution region. In order to avoid singularities, a Duffy transformation is applied, which allows singularity-free integration and maximized convergence region when performed with the help of Differential Algebraic methods. The method is shown to perform well on the examples studied. Practical implementation choices and some of their limitations are also explored.

SEMIABELIAN VARIETIES OVER SEPARABLY CLOSED FIELDS, MAXIMAL DIVISIBLE SUBGROUPS, AND EXACT SEQUENCES

Given a separably closed field $K$ of characteristic $p>0$ and finite degree of imperfection, we study the $\sharp$ functor which takes a semiabelian variety $G$ over $K$ to the maximal divisible subgroup of $G(K)$. Our main result is an example where $G^{\sharp }$, as a ‘type-definable group’ in $K$, does not have ‘relative Morley rank’, yielding a counterexample to a claim in Hrushovski [J. Amer. Math. Soc. 9 (1996), 667–690]. Our methods involve studying the question of the preservation of exact sequences by the $\sharp$ functor, and relating this to issues of descent as well as model-theoretic properties of $G^{\sharp }$. We mention some characteristic 0 analogues of these ‘exactness-descent’ results, where differential algebraic methods are more prominent. We also develop the notion of an iterative D-structure on a group scheme over an iterative Hasse field, which is interesting in its own right, as well as providing a uniform treatment of the characteristic 0 and characteristic $p$ cases of ‘exactness descent’.

Extended space method for parameter identifiability of DAE systems

Mathematical models of physical systems often have parameters that must be identified from physical data. This makes the analysis of the parameter identifiability of the given model system an essential prerequisite. Thus far, several methods have been proposed for analyzing the parameter identifiability of ordinary differential equation (ODE) systems. But, to the best of our knowledge, the parameter identifiability of differential algebraic equation (DAE) systems has scarcely been analyzed as a specific topic. Traditional differential algebraic (DA) methods developed for ODE systems are often applied directly on DAE systems. These methods, however, are not always applicable, e.g., when the prime ideal condition is not satisfied by a DAE system. In this paper, we propose a novel method to analyze the identifiability of DAE systems, based on the concept of space extension, through which the algebraic and differential variables can be decoupled. Furthermore, an inherent, low-dimensional, regular ODE system can be obtained, which is the external equivalent of the original DAE system. Subsequently, the differential algebraic (DA) method can then be used to analyze the identifiability of the low-dimension ODE system. Theoretical analysis is also presented for the proposed method. Two examples, including a simplified interaction model and an isothermal reactor system, are presented to illustrate the detailed steps and effectiveness of the proposed method.

Structural Conditions on Observability of Nonlinear Systems

In this paper parameter space and Lebesgue measurement are introduced into analysis of nonlinear systems. Structural observability rank condition is defined and together with the distinguishabililty the structural observability criterions of nonlinear systems are obtained. It proves that when the parameters are not identifiable the solutions with the same time but different parameters are also indistinguishable. Differential geometry and algebraic methods are used to investigate the observability problem, and it is proved that there are some relations between these two methods. Finally, examples are used to illustrate applications of the structural observability criterions.

The fast multipole method in the differential algebra framework

A method is presented that allows the computation of space charge effects of arbitrary and large distributions of particles in an efficient and accurate way based on a variant of the Fast Multipole Method (FMM). It relies on an automatic multigrid-based decomposition of charges in near and far regions and the use of high-order differential algebra methods to obtain decompositions of far fields that lead to an error that scales with a high power of the order. Given an ensemble of N particles, the method allows the computation of the self-fields of all particles on each other with a computational expense that scales as O(N). Using remainder-enhanced DA methods, it is also possible to obtain rigorous estimates of the errors of the methods. Furthermore, the method allows the computation of all high-order multipoles of the space charge fields that are necessary for the computation of high-order transfer maps and all resulting aberrations.

Optimal correction and design parameter search by modern methods of rigorous global optimization

Frequently the design of schemes for correction of aberrations or the determination of possible operating ranges for beamlines and cells in synchrotrons exhibit multitudes of possibilities for their correction, usually appearing in disconnected regions of parameter space which cannot be directly qualified by analytical means. In such cases, frequently an abundance of optimization runs are carried out, each of which determines a local minimum depending on the specific chosen initial conditions. Practical solutions are then obtained through an often extended interplay of experienced manual adjustment of certain suitable parameters and local searches by varying other parameters.However, in a formal sense this problem can be viewed as a global optimization problem, i.e. the determination of all solutions within a certain range of parameters that lead to a specific optimum. For example, it may be of interest to find all possible settings of multiple quadrupoles that can achieve imaging; or to find ahead of time all possible settings that achieve a particular tune; or to find all possible manners to adjust nonlinear parameters to achieve correction of high order aberrations. These tasks can easily be phrased in terms of such an optimization problem; but while mathematically this formulation is often straightforward, it has been common belief that it is of limited practical value since the resulting optimization problem cannot usually be solved.However, recent significant advances in modern methods of rigorous global optimization make these methods feasible for optics design for the first time. The key ideas of the method lie in an interplay of rigorous local underestimators of the objective functions, and by using the underestimators to rigorously iteratively eliminate regions that lie above already known upper bounds of the minima, in what is commonly known as a branch-and-bound approach.Recent enhancements of the Differential Algebraic methods used in particle optics for the computation of aberrations allow the determination of particularly sharp underestimators for large regions. As a consequence, the subsequent progressive pruning of the allowed search space as part of the optimization progresses is carried out particularly effectively. The end result is the rigorous determination of the single or multiple optimal solutions of the parameter optimization, regardless of their location, their number, and the starting values of optimization.The methods are particularly powerful if executed in interplay with genetic optimizers generating their new populations within the currently active unpruned space. Their current best guess provides rigorous upper bounds of the minima, which can then beneficially be used for better pruning. Examples of the method and its performance will be presented, including the determination of all operating points of desired tunes or chromaticities, etc. in storage ring lattices.

Differential algebraic methods for single particle dynamics studies of the University of Maryland electron ring

The University of Maryland electron ring is a small low energy machine for the study of space-charge dominated beams. Differential algebraic methods as implemented in COSY INFINITY offer an accurate method to study and analyze single particle nonlinear dynamics. As a starting point for space-charge related studies, we undertook a comprehensive examination of the single particle nonlinear dynamics based on differential algebra methods. Quantities such as tunes, chromaticities, dispersion, amplitude dependent tune shifts, and resonance strengths were calculated, and robustness of the solutions with respect to errors tested. The model demonstrated that the earth’s magnetic field has a significant impact on the beam, and adds rich dynamics even in the absence of space charge. Initially we determined the tunes for which an injection-free idealization of the ring had the largest dynamic aperture. Our study then showed that the actual ring also had the largest dynamic aperture at these same tunes, and at these tunes was also least sensitive to errors. Comparison of predicted beam trajectories with measured data showed that the model was accurate for the examined area.

Identification Aspects of Ritt's Algorithm for Discrete-Time Systems

In system identification, the challenge of parameter estimation often lies in solving a non-convex optimization problem. In many cases, this implies that it is difficult to guarantee that the global optimum will be found. The tools of differential algebra, for example, Gröbner bases and Ritt's algorithm, have turned out to be quite useful when dealing with certain nonlinear model structures. Some examples of successful applications are the determination of controllability, observability and global identifiability of these model structures. In this paper, difference algebraic techniques, which mimics the differential algebraic methods, will be presented. Besides making it possible to handle discrete-time systems, this opens up the possibility of dealing with noise. It turns out that the classical instrumental variables method plays a role.

Multiscale Representation and Segmentation of Hyperspectral Imagery Using Geometric Partial Differential Equations and Algebraic Multigrid Methods

A fast algorithm for multiscale representation and segmentation of hyperspectral imagery is introduced in this paper. The multiscale/scale-space representation is obtained by solving a nonlinear diffusion partial differential equation (PDE) for vector-valued images. We use algebraic multigrid techniques to obtain a fast and scalable solution of the PDE and to segment the hyperspectral image following the intrinsic multigrid structure. We test our algorithm on four standard hyperspectral images that represent different environments commonly found in remote sensing applications: agricultural, urban, mining, and marine. The experimental results show that the segmented images lead to better classification than using the original data directly, in spite of the use of simple similarity metrics and piecewise constant approximations obtained from the segmentation maps.

Long-term stability of the Tevatron by verified global optimization

The tools used to compute high-order transfer maps based on differential algebraic (DA) methods have recently been augmented by methods that also allow a rigorous computation of an interval bound for the remainder. In this paper we will show how such methods can also be used to determine rigorous bounds for the global extrema of functions in an efficient way. The method is used for the bounding of normal form defect functions, which allows rigorous stability estimates for repetitive particle accelerator. However, the method is also applicable to general lattice design problems and can enhance the commonly used local optimization with heuristic successive starting point modification. The global optimization approach studied rests on the ability of the method to suppress the so-called dependency problem common to validated computations, as well as effective polynomial bounding techniques. We review the linear dominated bounder (LDB) and the quadratic fast bounder (QFB) and study their performance for various example problems in global optimization. We observe that the method is superior to other global optimization approaches and can prove stability times similar to what is desired, without any need for expensive long-term tracking and in a fully rigorous way.