Extreme Value Theorem

According to this theorem, a function /(x) in a closed and bounded domain must attain minimum and maximum values. This theorem is also known as the Extreme Value Theorem. 

We now present two theorems which can be used to find the values of the time-domain function at two extremes, t = 0 and t = °°, without having to do the inverse transform. In control, we use the final value theorem quite often. The initial value theorem is less useful. As we have seen from our very first example in Section 2.1, the problems that we solve are defined to have exclusively zero initial conditions. 

GIBBS-KONOVALOV THEOREMS. Consider a binary system containing two phases (e.g.. liquid and vapor). Both components can pass from one phase lo another. The Gibbs Konovalov theorems refer to the properties of the phase diagrams of such systems (see also Azeotropic System). The lirst theorem is At constant pressure, the temperature of coexistence passes through tin extreme value (maximum, minimum or inflexion with a horizontal value), if the comfutsirlon of the two phases is the same. Conversely, al a point at winch the temperature passes through an extreme value, the phases have the same composition. The second theorem is similar. It refers lo the coexistence pressure at constant temperature. 

In the case of a set of isothermal equilibrium states we have already seen that if the pressure passes through an extreme value, then a sufficient condition for this is that the two phases shall have identical compositions. (Gibbs-Konovalow theorem). We now require to find the condition that this extreme value is a maximum or minimum, and to do this it is necessary to investigate the second differentials of the coexistence curves. 

We shall not repeat here any discussion of the properties already established for states of uniform composition. It will be recalled that a state of uniform composition corresponds to an extreme value (maximum, minimum, or inflexion with a horizontal tangent) of the equilibrium pressure at constant temperature, or of the equilibrium temperature at constant pressure (Gibbs-Konovalow theorems, chap. XVIII, 6 and 9). 

These two theorems are general and include as particular cases the theorems established in chap. XVIII, 6 and in chap. XXIII. They do not however apply to mono variant or invariant systems. Thus the eutectic point, which is certainly an indifferent point, does not represent, mathematically, an extreme value of or p for it is the point of intersection of two curves each of which refers to a two-phase system e.g. solution + ice or solution-f salt) under constant pressure. Only at the eutectic do three phases (solution + salt + ice) coexist. A mono variant three-phase system does not have an isobaric curve. 

Also note that the Hamilton and Lagrange densities differ only in terms that are not functions of the control variable. Their extreme values as functions of u only occur for the same value of u. Therefore, the optimum theorem can also be written as... 

Since the first derivative of X(t) on the interval [T, k N exists, X(t) is continuous on such interval. If X(t) is monotoimus in the interval [7., 7(. j], the extreme values of X(t) are reached at the boundary points of the interval. Otherwise, the extreme values of X(t) can be obtained by comparing all the local extreme values found by Fermat theorem (Bronshtein et al. 1985). To calculate the value of X(t), Runge-Kutta methods can be applied for the numerical solution of the ordinary differential equations (Hairer et al. 1993, Hairer et al. 1996). 

Of course, both statements can be proved from the theorem of uniqueness for the attraction field. In addition, it is appropriate to comment a linear function reaches its maximum at terminal points of the interval. The same behavior is observed in the case of harmonic functions, which cannot have their extreme inside the volume. Otherwise, the average value of the function at some point will not be equal to its value at this point, and, correspondingly, the Laplacian would differ from zero. At the same time, saddle points may exist. 

It should be clear that the most likely or physical rate of first entropy production is neither minimal nor maximal these would correspond to values of the heat flux of oc. The conventional first entropy does not provide any variational principle for heat flow, or for nonequilibrium dynamics more generally. This is consistent with the introductory remarks about the second law of equilibrium thermodynamics, Eq. (1), namely, that this law and the first entropy that in invokes are independent of time. In the literature one finds claims for both extreme theorems some claim that the rate of entropy production is... 

The prior distribution is often not what is observed, and there can be extreme deviations from it. [1, 3, 23] By our terminology this means that the dynamics do impose constraints on what can happen. How can one explicitly impose such constraints without a full dynamical computation At this point I appeal again to the reproducibility of the results of interest as ensured by the Monte Carlo theorem. The very reproducibility implies that much of the computed detail is not relevant to the results of interest. What one therefore seeks is the crudest possible division into cells in phase space that is consistent with the given values of the constraints. This distribution is known as one of maximal entropy". 

The ratio of the mean solution concentration to the feed salt concentration is plotted in Figure 2 as a function of volume flux. Two extreme cases will be of interest. First, when qd/5 - , the right-hand side of Eq, (32) approaches unity. This means that for all values of r, the mean solution concentration becomes the feed salt concentration as the volume flux becomes infinite. The second case is when qd/Dg is very small. In this case Eq. (32) becomes indeterminant. Hence, L Hospltal s Theorem must be applied to find the limiting value. It gives two different... 

The variation theorem has been an extremely powerful tool in quantum chemistry. One important technique made possible by the variation theorem is the expression of a wave function in terms of variables, the values of which are selected by minimizing the expectation value of the energy. 

Ht( y) = H(t ) = H (t ) / H( z) for complex values of the dilatation parameter jj, i.e., H( ) is non-Hermitian and as such, the variational theorem does not apply. However, there exists a bi-variational theorem /57,58/ for non-Hermitian operators. The bi-variational SCF equations for the dilated Hamiltonians are derived by extremizing the generalized functional... 

The reciprocity theorem shows that for any wavefield we can switch between the receiver and source positions without changing the values of the observed field. This result plays an extremely important role in wavefield imaging and inversion, especially in wavefield migration, which will be discussed in Chapter 15. 

Investigations of Braune and Koref.—We are indebted to Braune and Koref (77, and particularly 986) for an extremely careful and painstaking test of the application of the Heat Theorem to condensed systems. The test was made on a number of galvanic cells in exactly the same manner as that already employed by U. Fischer. In all the cells U was found to have nearly the same value, whether determined from the temperature coefficient of the potential or by direct thermochemical means this provides a guarantee that the cells under observation were actually controlled by the process assumed to be supplying the current, though as a matter of fact there was hardly any doubt of this in the cases examined. 

The values in the last column but one, calculated by means of the Heat Theorem, agree extremely well with the mean of Uj and Ua (last column but two) the differences, sometimes positive and sometimes negative, are all within the errors of observation which affect the values of Ux and U3, and also of the specific heats, and may easily amount to 200 cals, by accumulation. The mean value of U, U and U2 should give us the heat evolution in the re-spective reactions with an ac- > curacy hitherto unknown in such W cases. W... 

Does it make sense to associate a definite quantum number n. to each mode / in an anharmonic system In general, this is an extremely difficult question But remember that so far, we are speaking of the situation in some small vicinity of a minimum on the PES, where the Moser-Weinstein theorem guarantees the existence of the anharmonic normal modes. This essentially guarantees that quantum levels with low enoughvalues correspond to trajectories that lie on invariant tori. Since the levels are quantized, these must be special tori, each characterized by quantized values of the classical actions I. = n. + 5)/), which are constants of the motion on the invariant toms. As we shall see, the possibility of assigning a set of iV quantum numbers n- to a level, one for each mode, is a very special situation that holds only near the potential minimum, where the motion is described by the N anharmonic normal modes. However, let us continue for now with the region of the spectmm where this special situation applies. 

It is thus not possible to obtain a phase-sensitive 45 projection of a normal homonuclear 2D J spectrum, and instead it is usual to project the absolute value or the power spectrum. This gives rise to distorted intensities and peak positions, but nevertheless is an extremely useful aid to assignment. The subject of phase-sensitive proton spectra without multiplet splittings is not however entirely closed, despite the clear message of the projection-cross-section theorem, and will be returned to later. In the interim, some of the many data acquisition and handling methods discussed above will be illustrated with some experimental spectra. 

The extreme action effects of structures are caused by service and chmate loads and may be modeled as intermittent rectangular renewal pulse processes. Therefore, it is e q)edient to treat the safety margin of particular members as a random sequence. The revised values of instantaneous survival probabihty of particular members may be analyzed by the concepts of truncated probabihty distribution and Bayes theorem. The presented new design methodology based on conventional resistances, rank sequences, correlation factors and transformed conditional probabihties may be successfully used in the prediction of long-term survival probabilities of members and their systems during residual service Ufe. 

The statistical properties of the random force f(0 are modeled with an extreme economy of assumptions f(t) is assumed to be a stationary and Gaussian stochastic process, with zero mean (f(0 = 0), uncorrelated with the initial value v(t = 0) of the velocity fluctuations, and delta-correlated with itself, f(0f(t ) = f25(t -1 ) (i.e it is a purely random, or white, noise). The stationarity condition is in reality equivalent to the fluctuation-dissipation relation between the random and the dissipative forces in Equation 1.1, which essentially fixes the value of y. In fact, from Equation 1.1 and the assumed properties of f(t), we can derive the expression y(f)v(t) = exp [v(0)v(0) -+ ylM °, where Xg = In equilibrium, the long-time asymptotic value y/M must coincide with the equilibrium average (vv) = (k TIM)t given by the equipartition theorem (with I being the 3 X 3 Cartesian unit tensor), and this fixes the value of y to y=... 

---