Surrogate Hamiltonian Monte Carlo uses an auxiliary smooth probability density function $\nu$ to generate trajectories but applies a Metropolis–Hastings acceptance probability using the intended target $\mu$. For Gaussian momenta and an involutive Metropolis proposal generated by flipped leapfrog integration, the acceptance probability is $\min (1,\mu (x^{\prime })e^{-\Vert p^{\prime }{\Vert }^2/2}/[\mu (x)e^{-\Vert p{\Vert }^2/2}])$. Only the surrogate gradient is required during the trajectory; target density evaluations are still required at the endpoints.

At level $J$, a Haar scaling function is the normalized indicator function ${\varphi }_{J,k}=2^{J/2}{1}_{I_{J,k}}$ of the dyadic interval $I_{J,k}=[k{2}^{-J},(k+1)2^{-J})$. Its L2 norm is one because its squared height is $2^J$ and its support length is $2^{-J}$.

For a real square-integrable function $f$, its coefficient isMultiplying the coefficient by the Haar scaling function restores the cell average on that interval. The disjoint supports make the functions at one level an orthonormal basis of the cellwise-constant subspace. This normalization is also what turns coefficient variances in the Gaussian white noise model into $1/n$ and cellwise variances into $2^J/n$ in the Haar projection estimator in Gaussian white noise.

The Haar projection at level $J$ is the orthogonal projection onto the functions constant on level-$J$ dyadic intervals. On each cell it replaces $f$ by its cell average. This is the closest cellwise-constant approximation in the L2 norm, because choosing a constant $c$ on a cell minimizes ${\int }_I|f-c{|}^2$ at its cell average.

On $[0,1]$, take the step function with values $2,4,6,8$ on successive quarter-intervals. At level zero, its Haar projection is the constant $5$. At level one, it has values $3$ and $7$ on the two halves. At level two, it recovers the four original values exactly. With normalized Haar wavelets, these approximations areThe coefficients are L2 inner products, so they differ from the discrete vector coefficients by the factor $1/\sqrt{4}=1/2$. Their squared sum also has a direct meaning:The approximation errors in squared L2 norm of $P_{0}f,P_{1}f,P_{2}f$ are respectively $5,1,0$: each added detail removes exactly its own squared coefficient from the error.

The Haar projection can be expressed using level-$J$ Haar scaling functions, or through a coarse scaling family and the Haar wavelets at intermediate levels. On $[0,1]$ with lowest level zero, that coarse family is the single constant function $\varphi$. On the whole real line, use the integer translates of a coarse Haar scaling function instead; one global constant is not square-integrable. Refinement improves the representation while retaining the earlier coefficients.

For equal particle rates in the GRW model, the next centre falls in $B$ with probability $N^{-1}{\sum }_i{\int }_B\Vert L_i(c)\Psi {\Vert }^{2}dc$. With zero Hamiltonian operator, two successive centres in $B$ have joint probability $N^{-2}{\sum }_{ij}{\int }_{B}dc{\int }_{B}dd\Vert L_j(d)L_i(c)\Psi {\Vert }^2$. Dividing by the first-event probability gives the conditional second event, including repeated particle labels. Spatial probabilities depend on actual packet mass, not solely the nominal packet centres.

A first localization near one macroscopically separated branch suppresses the other branch through the Gaussian GRW localization operator. This selects a correlated pointer alternative. However, if a region contains only packet centres and excludes substantial packet mass, the next centre need not lie in that region's small enlargement. Probabilities $|\alpha {|}^2$ for the first event and approximately one for the next require regions containing essentially all corresponding branch mass. The general GRW collapse-centre event probabilities retain the geometry exactly.

In an accepted padded convolution of words, a loop after one tape ends can be deleted or repeated while keeping that entire tape fixed. The altered other tape still represents the same required neighbouring group element. A shortest representative has a suffix no longer than the fixed number of automaton states; an accepted longer suffix gives infinitely many representatives of the same element.

Fix the first tape of a group multiplier automaton and use dynamic programming over automaton states and output-padding status, retaining predecessor pointers. The pumping of a group multiplier automaton bounds the search by $|u|+N$ layers, each of fixed size. This constructs a representative of $\overline{u}x$ in time $O(|u|+1)$ without assuming uniqueness of representatives.

For a two-dimensional Newtonian fluid ice sheet attached to an unbuttressed ice shelf, ice-sheet flotation gives $\rho H={\rho }_{w}b$. The difference between the depth-integrated ice and ocean hydrostatic pressures is $\rho g{H}^2/2-{\rho }_{w}g{b}^2/2=\rho g^{\prime }{H}^2/2$, where $g^{\prime }=g(1-\rho /{\rho }_w)$. Shelf force balance integrates $4\mu H{u}_x-\rho g^{\prime }{H}^2/2$ to a constant; absence of a buttressing force makes that constant zero. The shallow-shelf approximation therefore gives the displayed boundary condition, or $u_x=g^{\prime }H/(8\nu )$ with kinematic viscosity $\nu =\mu /\rho$.

For a bed depth $b=\alpha x$, a thin subglacial till layer of thickness $l$ is idealized by Newtonian till lubrication, with dynamic viscosity $\lambda \mu$ and basal drag $\lambda \mu u/l$. Neglecting longitudinal stress divergence in the bulk yields $u=-(l/\lambda )(g/\nu )H{h}_x$. Retaining the unbuttressed Newtonian grounding-line stress condition as boundary data and differentiating ice-sheet flotation along the moving grounding line givesIndeed, mass conservation gives $H_t=-u{H}_x-H{u}_x$, $H_x=h_x+\alpha$, and ice-sheet flotation gives $H_t=(\alpha {\rho }_w/\rho -H_x){\dot{x}}_G$. This is a formal local friction closure, not a uniformly valid removal of membrane stress near every grounding line. If the coefficient of ${\dot{x}}_G$ vanishes, it is an implicit compatibility condition and cannot be divided to obtain a finite speed.

At a grounding line, hydrostatic pressures balance when the weight of an ice column of thickness $H$ and mass density $\rho$ equals the weight of its displaced water, of depth $b$ and mass density ${\rho }_w$. Thus flotation gives $\rho H={\rho }_{w}b$. Its time derivative along a moving grounding line relates local thickness changes to grounding-line speed; it supplies a different boundary condition from the unbuttressed Newtonian grounding-line stress.

The healing length of a repulsive homogeneous condensate is the length at which density-gradient kinetic energy and interaction energy balance. The displayed convention sets the dimensionless Laplacian coefficient to one in the stationary Gross–Pitaevskii equation. A hard-wall number density profile has width $\sqrt{2}{\ell }_h$; definitions without the factor $\sqrt{2}$ occur, so the equation and convention must be specified.

For the normalized conservative equation $-2i{\psi }_t={\nabla }^2\psi +(1-|\psi {|}^2)\psi$, a localized disturbance translating at speed $v$ obeys ${\nabla }^2\psi +(1-|\psi {|}^2)\psi -2iv{\psi }_x=0$ in translating coordinates. Its complex-field boundary is $\psi \to 1$. This permits zero-net-winding vortex pairs, but excludes a lone quantum vortex. The usual subsonic far field has $\psi -1=O(r^{-1})$ in two dimensions.

Neglecting density-gradient kinetic energy in the stationary Gross–Pitaevskii equation gives $n=(\mu -V{)}_{+}/g$ for repulsive interactions. For a radial harmonic trap $V=m{\omega }_r^2{r}^2/2$, this is $n=n_0(1-r^2/R_r^2{)}_{+}$ with $R_r=\sqrt{2g{n}_0/(m{\omega }_r^2)}$. Hard boundaries can truncate this natural support. The approximation fails in vortex cores and boundary healing regions and requires a cloud much larger than its healing length.

For $i{\psi }_t=-D{\nabla }^2\psi +[V(n)+i\Gamma (n)]\psi$, with real $D,V,\Gamma$ and $\psi =\sqrt{n}{e}^{iS}$, the continuity equation isMultiply the equation by ${\psi }^{\ast }$ and subtract its complex conjugate multiplied by $\psi$ to obtain this identity; real interaction and potential terms cancel. Thus the current velocity is $\mathbf{j}/n=2D\nabla S$. With dimensional kinetic operator $-{\hslash }^2{\nabla }^2/(2m)$ and left side $i\hslash {\psi }_t$, this gives the physical superfluid velocity $(\hslash /m)\nabla S$. In normalized equations with $D=1/2$ or $D=1$, the current velocity is respectively $\nabla S$ or $2\nabla S$. The distinction fixes the radial-flow slope of a driven quantum vortex.

For a local two-body hyperbolic encounter with central parameter $\mu$ and allowed centre separation $r_{\min }\ge r_0$, conservation of energy gives equal asymptotic speeds. The gravitational scattering angle therefore impliesPut $z=v_{\infty }\sqrt{r_{\min }/\mu }$ and use $2z/(1+z^2)\le 1$. The optimizing grazing orbit has $z=1$, orbital eccentricity two and deflection $\pi /3$. For finite spherical bodies $r_0$ is the sum of their radii; excluding actual contact makes the upper bound a supremum. Without a finite exclusion radius the ideal point-mass problem has no finite bound.

A Grassmann derivative is an odd differentiation operation on a Grassmann algebra. A left Grassmann derivative obeys ${\partial }_{\theta }^L(ab)=({\partial }_{\theta }^{L}a)b+(-1{)}^{|a|}a({\partial }_{\theta }^{L}b)$ for homogeneous $a$. Right derivatives use the corresponding right-sided product rule. Their placement must be fixed when differentiating fermionic sources in a generating functional, since commuting ordinary derivatives would lose fermionic signs.

The limiting coderivative is $D^{\ast }S(\bar{u}\mid \bar{v})(v^{\ast })=\{u^{\ast }:(u^{\ast },-v^{\ast })\in N_{\mathrm{gph}S}(\bar{u},\bar{v})\}$. The negative sign on the output dual is essential. For a smooth single-valued function it gives the transpose Jacobian acting on $v^{\ast }$.

The graded shift $M(-a)$ is the same underlying module with $M(-a{)}_d=M_{d-a}$. Thus $S(-a)$ has a free generator in degree $a$. Its Hilbert series is $t^a{H}_M(t)$, and multiplication by a homogeneous element of degree $a$ becomes a degree-preserving map $M(-a)\to M$.

If $f$ is homogeneous of degree one, multiplication by powers of the invertible element $f$ identifies every homogeneous component of $M_f$ with $(M_f{)}_0$. A homogeneous element $m$ of degree $a$ is written $f^a(f^{-a}m)$. Thus the displayed map is an isomorphism of graded modules. In particular, the degree-zero part of a tensor product of localized graded modules is the tensor product of their degree-zero parts over $(S_f{)}_0$. This proves tensor compatibility of graded sheafification on degree-one-generated Proj.

For finite free total chain complexes, evaluation identifies the underlying graded tensor product $X\otimes C^{\prime }$ with the graded Hom complex of chain complexes, where $X$ is the reversed dual chain complex. Under the precomposition-first Hom differential, conjugation by the sign $(-1{)}^{j(j-1)/2}$ makes this a chain isomorphism. The identity $\rho (j)-\rho (j-1)=j-1$ works for every integer degree. If the grading is unbounded and only degreewise finite, products on the Hom side need not equal direct sums on the tensor side.