Dynamical systems

Dynamical systems play a pivotal role across various scientific domains, encompassing disciplines from physics to biology and engineering. The long-term behavior of these systems hinges on the structure of their attractors, with many exhibiting multistability characterized by multiple minimal attractors. Understanding the structure of these attractors and their corresponding basins is a central theme in dynamical systems theory.
In recent years, machine learning algorithms have emerged as potent tools for clustering, prediction, and modeling complex data. By harnessing the capabilities of neural networks along with techniques from topological data analysis, in particular persistence homology, we can construct surrogate models of system asymptotics. This approach also allows for the decomposition of phase space into polygonal regions and the identification of plausible attracting neighborhoods, facilitating homological Conley index computation at reduced computational expense compared to current methods. Through various illustrative examples, we demonstrate that sufficiently low training loss yields constructed neighborhoods whose homological Conley indices aligns with a priori knowledge of the dynamics.

We explore a novel method of approximating contractible invariant circles in maps. The process begins by leveraging improvements on Birkhoff's Ergodic Theorem via Weighted Birkhoff Averages to compute high precision estimates on several Fourier modes. We then set up a Newton-like iteration scheme to further improve the estimation and extend the approximation out to a sufficient number of modes to yield a significant decay in the magnitude of the coefficients of high order. With this approximation in hand, we explore the phase space near the approximate invariant circle with a form numerical continuation where the rotation number is perturbed and the process is repeated. Then, we turn our attention to a completely different problem which can be approached in a similar way to the numerical continuation, finding a Siegel disk boundary in a holomorphic map. Given a holomorphic map which leads to a formally solvable cohomological equation near the origin, we use a numerical continuation style process to approximate an invariant circle in the Siegel disk near the origin. Using an iterative scheme, we then enlarge the invariant circle so that it approximates the boundary of the Siegel disk.

In the past few years, the development of complex dynamical networks or systems has stimulated great interest in the study of the principles and mechanisms underlying the Internet of things (IoT). IoT is envisioned as an intelligent network infrastructure with a vast number of ubiquitous smart devices present in diverse application domains and have already improved many aspects of daily life. Many overtly futuristic IoT applications acquire data gathered via distributed sensors that can be uniquely identified, localized, and communicated with, i.e., the support of sensor networks. Soft-sensing models are in demand to support IoT applications to achieve the maximal exploitation of transforming the information of measurements into more useful knowledge, which plays essential roles in condition monitoring, quality prediction, smooth control, and many other essential aspects of complex dynamical systems. This in turn calls for innovative soft-sensing models that account for scalability, heterogeneity, adaptivity, and robustness to unpredictable uncertainties. The advent of big data, the advantages of ever-evolving deep learning (DL) techniques (where models use multiple layers to extract multi-levels of feature representations progressively), as well as ever-increasing processing power in hardware, has triggered a proliferation of research that applies DL to soft-sensing models. However, many critical questions need to be further investigated in the deep learning-based soft-sensing.

This dissertation is a study about applied dynamical systems on two concentrations. First, on the basis of the growing association between opioid addiction and HIV infection, a compartmental model is developed to study dynamics and optimal control of two epidemics; opioid addiction and HIV infection. We show that the disease-free-equilibrium is locally asymptotically stable when the basic reproduction number R0 = max(Ru0; Rv0) < 1; here Rv0 is the reproduction number of the HIV infection, and Ru0 is the reproduction number of the opioid addiction. The addiction-only boundary equilibrium exists when Ru0 > 1 and it is locally asymptotically stable when the invasion number of the opioid addiction is Ruinv < 1: Similarly, HIV-only boundary equilibrium exists when Rv0 > 1 and it is locally asymptotically stable when the invasion number of the HIV infection is Rvinv < 1. We study structural identifiability of the parameters, estimate parameters employing yearly reported data from Central for Disease Control and Prevention (CDC), and study practical identifiability of estimated parameters. We observe the basic reproduction number R0 using the parameters. Next, we introduce four distinct controls in the model for the sake of control approach, including treatment for addictions, health care education about not sharing syringes, highly active anti-retroviral therapy (HAART), and rehab treatment for opiate addicts who are HIV infected. US population using CDC data, first applying a single control in the model and observing the results, we better understand the influence of individual control. After completing each of the four applications, we apply them together at the same time in the model and compare the outcomes using different control bounds and state variable weights. We conclude the results by presenting several graphs.

The goal of this work is to study smooth invariant sets using high order approximation schemes. Whenever possible, existence of invariant sets are established using computer-assisted proofs. This provides a new set of tools for mathematically rigorous analysis of the invariant objects. The dissertation focuses on application of these tools to a family of three dimensional dissipative vector fields, derived from the normal form of a cusp-Hopf bifurcation. The vector field displays a Neimark-Sacker bifurcation giving rise to an attracting invariant torus. We examine the torus via parameter continuation from its appearance to its breakdown, scrutinizing its dynamics between these events. We also study the embeddings of the stable/unstable manifolds of the hyperbolic equilibrium solutions over this parameter range. We focus on the role of the invariant manifolds as transport barriers and their participation in global bifurcations. We then study the existence and regularity properties for attracting invariant tori in three dimensional dissipative systems of ordinary differential equations and lay out a constructive method of computer assisted proof which pertains to explicit problems in non-perturbative regimes. We get verifiable lower bounds on the regularity of the attractor in terms of the ratio of the expansion rate on the torus with the contraction rate near the torus. We look at two important cases of rotational and resonant tori. Finally, we study the related problem of approximating two dimensional subcenter manifolds of conservative systems. As an application, we compare two methods for computing the Taylor series expansion of the graph of the subcenter manifold near a saddle-center equilibrium solution of a Hamiltonian system.

Gravitational N-body problems are central in classical mathematical physics.
Studying their long time behavior raises subtle questions about the interplay between
regular and irregular motions and the boundary between integrable and chaotic dynamics.
Over the last hundred years, concepts from the qualitative theory of dynamical
systems such as stable/unstable manifolds, homoclinic and heteroclinic tangles,
KAM theory, and whiskered invariant tori, have come to play an increasingly important
role in the discussion. In the last fty years the study of numerical methods for
computing invariant objects has matured into a thriving sub-discipline. This growth
is driven at least in part by the needs of the world's space programs.
Recent work on validated numerical methods has begun to unify the computational
and analytical perspectives, enriching both aspects of the subject. Many
of these results use computer assisted proofs, a tool which has become increasingly
popular in recent years. This thesis presents a proof that the circular restricted four
body problem is non-integrable. The proof of this result is obtained as an application
of more general rigorous numerical methods in nonlinear analysis.