Continuity Study Guide

📖 Core Concepts Continuity (math) – opposite of discreteness; a property that lets a function be drawn without lifting the pen. Continuous function – for every small change in input, the output changes only a small amount (no jumps). Topological continuity – the same “no‑jump” idea but defined for maps between abstract spaces using open‑set pre‑images. Parametric continuity (Cⁿ) – smoothness of a curve measured by matching derivatives up to order n (e.g., \(C^1\) = tangent continuity, \(C^2\) = curvature continuity). Geometric continuity (Gⁿ) – visual smoothness of a shape; only the shape of the curve matters, not the parameter speed (e.g., \(G^1\) matches tangents, \(G^2\) matches curvature direction). Higher‑order continuity – \(C^2\) or \(G^2\) ensures curvature continuity, giving aesthetically pleasing surfaces in CAD/animation. Continuous probability distribution – a random variable can take any value in an interval; its probability is described by a density function. Continuous stochastic process – each sample path is a continuous function of time (e.g., Brownian motion). Continuity equations (physics) – mathematical statements of conserved quantities (mass, energy, momentum, electric charge, probability) expressed as \(\frac{\partial \rho}{\partial t} + \nabla\!\cdot\! \mathbf{J}=0\). --- 📌 Must Remember Lévy’s continuity theorem – convergence in distribution of random variables ⇔ pointwise convergence of characteristic functions. \(C^n\) vs. \(G^n\) – \(C^n\) requires matching derivatives; \(G^n\) only requires matching geometric features (direction of tangent, curvature). Continuity equation form – \(\displaystyle \frac{\partial \rho}{\partial t} + \nabla\!\cdot\!\mathbf{J}=0\) (ρ = density, J = flux). Key examples of continuous processes – Brownian motion, Ornstein‑Uhlenbeck process. Conserved quantity → continuity equation – mass → fluid flow, energy → thermodynamics, momentum → mechanics, charge → electromagnetism, probability → stochastic systems. --- 🔄 Key Processes Checking continuity of a real‑valued function Choose any point a. Verify \(\lim{x\to a} f(x) = f(a)\). If true for all a in the domain → function is continuous. Establishing \(C^n\) continuity between curve segments Align end‑point positions. Match first‑order derivatives (tangents) for \(C^1\). Match second‑order derivatives (curvature) for \(C^2\). Continue up to desired order n. Applying Lévy’s continuity theorem Compute characteristic functions \(\phi{Xk}(t)=E[e^{itXk}]\). Show \(\phi{Xk}(t) \to \phi(t)\) pointwise for all t. Conclude \(Xk \xrightarrow{d} X\) where \(\phi\) is the limit characteristic function. Deriving a continuity equation (e.g., mass) Write mass balance for a control volume: rate of accumulation = inflow − outflow. Convert inflow/outflow to flux divergence → \(\frac{\partial \rho}{\partial t} + \nabla\!\cdot\!\mathbf{J}=0\). --- 🔍 Key Comparisons \(C^n\) vs. \(G^n\) – \(C^n\): parameter‑dependent derivative matching. \(G^n\): parameter‑independent geometric matching (only direction/shape). Continuous random variable vs. discrete random variable – Continuous: values form an interval; probability expressed via density \(f(x)\). Discrete: values are isolated points; probability via mass function \(p(x)\). Continuity theorem (Lévy) vs. pointwise convergence – Lévy: convergence of distributions ⇔ convergence of characteristic functions (global). Pointwise: only checks function values at each point; not enough for distribution convergence. --- ⚠️ Common Misunderstandings “Continuous = smooth” – continuity allows kinks; smoothness requires at least \(C^1\). \(G^1\) guarantees equal tangents – actually guarantees parallel tangents; speed may differ. All stochastic processes with continuous paths are deterministic – false; Brownian motion is random yet continuous. Continuity equation only for fluids – it’s a universal form for any conserved scalar (mass, energy, charge, probability). --- 🧠 Mental Models / Intuition Pen‑without‑lifting – imagine drawing the graph; if you never lift the pen, the function is continuous. Road‑smoothness analogy – \(C^1\) = no sudden direction change; \(C^2\) = no sudden curvature change (no “bumps”). Conservation as “stuff can’t disappear” – continuity equation = “what comes in must either go out or accumulate”. Characteristic function as “frequency fingerprint” – if fingerprints converge, the underlying distributions converge (Lévy). --- 🚩 Exceptions & Edge Cases Functions continuous on an interval but not differentiable (e.g., \(f(x)=|x|\) at 0). Geometric continuity can be achieved with mismatched parameter speeds – still \(G^1\) even if \(C^1\) fails. Probability continuity – a distribution can have a density except at isolated points (mixed continuous‑discrete). Lévy’s theorem requires characteristic functions to be uniformly bounded – pathological cases may violate assumptions. --- 📍 When to Use Which Designing CAD surfaces – prefer \(G^2\) for visual smoothness; use \(C^2\) only when derivative control matters (e.g., physical simulation). Proving distribution convergence – apply Lévy’s continuity theorem when characteristic functions are easier to handle than PDFs/CDFs. Modeling physical conservation – write a continuity equation whenever a scalar quantity is conserved (mass, energy, charge, probability). Analyzing stochastic processes – check sample‑path continuity for Brownian motion; use SDEs for continuous‑time models. --- 👀 Patterns to Recognize “Derivative matching up to order n” → signals a \(C^n\) continuity requirement. Flux divergence term \(\nabla\!\cdot\!\mathbf{J}\) appearing with a time derivative → classic continuity equation pattern. Characteristic function limit appearing in probability questions → likely a Lévy continuity theorem cue. “No jumps” language → points to ordinary continuity (not necessarily smooth). --- 🗂️ Exam Traps Choosing \(C^1\) when only \(G^1\) is needed – you may over‑constrain a design, wasting parameters. Assuming a continuous random variable has a PDF everywhere – mixed distributions break this. Confusing “conserved quantity” with “constant quantity” – continuity equation allows local change via flux, not global constancy. Selecting pointwise convergence as proof of distribution convergence – ignores Lévy’s requirement on characteristic functions. Mistaking “smooth” for “continuous” – a function can be continuous but not differentiable; watch for kink questions.