Valuations and Surface Area Measures


 Ethan Dickerson
 2 years ago
 Views:
Transcription
1 Valuations and Surface Area Measures Christoph Haberl and Lukas Parapatits Abstract We consider valuations defined on polytopes containing the origin which have measures on the sphere as values. We show that the classical surface area measure is essentially the only such valuation which is SL(n) contravariant of degree one. Moreover, for all real p, an L p version of the above result is established for GL(n) contravariant valuations of degree p. This provides a characterization of the L p surface area measures from the L p BrunnMinkowski theory. Mathematics Subject Classification: 52B45, 28A0 Introduction A valuation is a map µ : Q n A, + defined on a set Q n of subsets of R n with values in an abelian semigroup such that µ(p Q) + µ(p Q) = µ(p ) + µ(q) whenever the sets P, Q, P Q, P Q are contained in Q n. Valuations were the critical ingredient in Dehn s solution of Hilbert s Third Problem and played a prominent role in geometry thereafter. Especially over recent years the theory of valuations witnessed an explosive growth (see e.g. [2 6, 9, 3 5, 20 25, 34, 39 4]). For example, it turned out that basic objects in convex geometric analysis can be characterized as valuations which are compatible with a subgroup of the affine group. Moreover, new geometric insights gained from such classification results led to strengthenings of various affine isoperimetric and Sobolev inequalities (see [8, 7, 8, 28, 30]). In this paper we classify measure valued valuations which are compatible with the general linear group. We will show that the only nontrivial examples of such valuations are surface area measures and their L p analogs. Surface area measures are a fundamental concept in the theory of convex bodies, i.e. nonempty compact convex subsets of R n. Given a convex polytope P R n, its surface area measure S(P, ) is the Borel measure on the Euclidean unit sphere S n R n
2 which is given by () S(P, ) = u N (P ) V n (F u )δ u. Here, N (P ) denotes the set of all unit facet normals of P, V n (F u ) stands for the (n )dimensional volume of the facet corresponding to u, and δ u is the Dirac probability measure supported at u. Surface area measures can actually be associated with arbitrary convex bodies. We write K n for the space of convex bodies in R n equipped with Hausdorff distance. For a body K K n, its surface area measure S(K, ) is simply the weak limit of the measures S(P k, ), where P k is some sequence of polytopes converging to K. Surface area measures are the central object of a cornerstone of the classical Brunn Minkowski theory: The Minkowski problem. It asks which measures on the Euclidean unit sphere are surface area measures of convex bodies. The answer to this question had a huge impact on convex geometry, geometric tomography, differential geometry, and elliptic partial differential equations (see e.g. [, 35, 38]). The first main result of this paper is a characterization of the surface area measure. We will consider valuations which are defined on polytopes containing the origin and which have measures on the sphere as values. We show that the surface area measure is essentially the only such valuation with a certain compability with the special linear group. What we mean by compatibility is contained in the following definition. Let G be a subgroup of the general linear group GL(n) and denote by M(S n ) the space of finite positive Borel measures on S n. Suppose that p R. A map µ : Q n M(S n ) defined on Q n K n is called G contravariant of degree p if f dµ(φp, ) = det φ f φ t dµ(p, ) S n S n for every map φ G, each P Q n with φp Q n, and every continuous phomogeneous function f : R n \{o} R. Here, φ t denotes the transpose of the inverse of φ GL(n). We remark that the concept of G contravariance is simply the behavior of mixed volumes (and their L p analogs) under the action of the general linear group (see Section 2). We are now in a position to state our first main result. Throughout this article we work in R n with n 3. Convex polytopes in R n which contain the origin are denoted by Po n. Theorem. A map µ : P n o M(S n ) is an SL(n) contravariant valuation of degree if and only if there exist constants c, c 2, c 3, c 4 R with c, c 2 0 and c + c 3 0, c 2 + c 4 0 such that µ(p, ) = c S(P, ) + c 2 S( P, ) + c 3 S (P, ) + c 4 S ( P, ) 2
3 for every P P n o. The measure S (P, ) is defined similarly to the surface area measure of P but the summation in () ranges only over those facets in N (P ) which do not contain the origin. An immediate consequence of the above result is the following classification of measure valued valuations defined on all convex bodies. Corollary 2. A map µ : K n M(S n ) is a weakly continuous, translation invariant, and SL(n) contravariant valuation of degree if and only if there exist nonnegative constants c, c 2 R such that for every K K n. µ(k, ) = c S(K, ) + c 2 S( K, ) Schneider [37] previously obtained a classification of rotation contravariant valuations of degree under the additional assumption that they are defined locally. As explained before, surface area measures lie at the very core of the Brunn Minkowski theory. Based on Firey s L p addition for convex bodies, Lutwak [26, 27] showed that the classical BrunnMinkowski theory can be extended to an L p Brunn Minkowski theory. The importance of this new L p theory is reflected for example in the fact that L p inequalities almost invariably turn out to be stronger than their classical counterparts. Since Lutwak s seminal work, this L p BrunnMinkowski theory evolved enormously (see e.g. [7, 0, 7, 8, 22, 25, 28, 29, 36, 42 45]). Let p R. The analog of the surface area measure in the L p BrunnMinkowski theory is defined as follows. For a convex polytope P Po n, the L p surface area measure S p (P, ) M(S n ) is given by S p (P, ) = h(p, u) p V n (F u )δ u, u N (P ) where N (P ) denotes the set of unit facet normals of P corresponding to facets which do not contain the origin and h(p, ) is the support function of P (see Section 2 for the precise definition). Finding necessary and sufficient conditions for a measure to be the L p surface area measure of a convex body is one of the major problems in modern convex geometric analysis. Consequently, this L p analog of the Minkowski problem has been studied intensively (see e.g. [7,6,26,43]). Solutions to the L p Minkowski problem were crucial for the proofs of affine versions of the PólyaSzegö principle and new affine Sobolev inequalities (see [8, 7, 8, 28, 30]). The following theorem provides a characterization of L p surface area measures for all p. 3
4 Theorem 3. Let p R\{}. A map µ : Po n M(S n ) is a GL(n) contravariant valuation of degree p if and only if there exist nonnegative constants c, c 2 R such that µ(p, ) = c S p (P, ) + c 2 S p ( P, ) for every P P n o. Since the L p Brunn Minkowski theory is based on an addition which makes sense only for p, most of the L p concepts are restricted to such p s. However, the above theorem reveals that the concept of L p surface area measures is independent of p in a very natural way. For positive p, we will actually prove a stronger version of Theorem 3. It will be shown in Theorem 7 that for such p s the L p surface area measure can actually be characterized as an SL(n) contravariant valuation. Moreover, Theorem 4 will show that L p surface area measures are characterized as valuations which are SL(n) contravariant of degree p for all p R provided that their images are discrete. Recently, the next step in the evolution of the BrunnMinkowski theory towards an Orlicz BrunnMinkowski theory has been made (see e.g. [6, 25, 3, 32]). Whereas some elements of the L p BrunnMinkowski theory have been generalized to an Orlicz setting, the Orlicz analog of the surface area measure is still unknown. This question actually motivated the axiomatic characterization of L p surface area measures obtained in this article. Since characterizing properties of L p surface area measures are now identified, they can possibly lead to the correct notion of Orlicz surface area measures. 2 Notation and Preliminaries In this section we collect the necessary definitions and facts about convex bodies. Excellent references for the theory of convex bodies are the books by Gardner [], Gruber [2], and Schneider [38]. We write R + for the set of positive real numbers. Given two vectors x, y R n we write x y for their standard Euclidean product. The Euclidean length of a vector x R n is denoted by x. If x R n is not equal to the zero vector, then we set x = x x. The canonical basis vectors of R n are denoted by e,..., e n. The standard simplex T n R n is the convex hull of the origin and the canonical basis vectors e,..., e n. We denote by T the intersection T n e where e stands for the hyperplane through the origin orthogonal to e. If p is positive, then we write C p + (R n ) for the space of nonnegative, continuous, phomogeneous functions from R n to R. 4
5 For the definition of L p surface area measures we already used the notion of support functions. The precise definition is as follows. Given a convex body K K n, its support function is defined as h(k, x) = max{x y : y K}, x R n. It follows from the inclusionexclusion principle that a valuation µ : P n o M(S n ) is uniquely determined by its values on ndimensional simplices having one vertex at the origin and its value on {o} (see [33] for a short proof). If µ is in addition SL(n) contravariant of degree p, then the uniqueness part of Riesz s representation theorem implies that beside its behavior at the origin µ is uniquely determined by its values on the simplices st n with s > 0. We summarize this in the following lemma. Lemma 4. Let p R. A valuation µ : P n o M(S n ) which is SL(n) contravariant of degree p is uniquely determined by its values on positive multiples of the standard simplex T n and its value on {o}. A measure µ M(S n ) is called continuous if singletons have µmeasure zero. We call it discrete, if there exists a countable set N S n such that µ(s n \N) = 0. The set of all discrete members of M(S n ) is denoted by M d (S n ). Note that for µ, ν M d (S n ) we have (2) µ = ν µ(x) = ν(x) for all x S n. Here we used the convention µ(x) := µ({x}) for x S n. For every µ M(S n ) there exists a unique pair consisting of a continuous measure µ c and a discrete measure µ d such that (3) µ = µ c + µ d. Let µ : Po n M(S n ) be given. If µ is a valuation, so is µ d. Indeed, since the valuation property has to be checked only for points by (2), the assertion directly follows from the decomposition (3). Note that if µ is SL(n) contravariant of degree p, then by the uniqueness part of Riesz s representation theorem and the transformation behavior of image measures we have fdµ(φp, ) = S n f φ t dµ(p, ) S n for all phomogeneous extensions of bounded Borel measurable functions f : S n R. In particular, the last relation holds for indicator functions of points. This yields (4) µ(φp, x ) x p = µ(p, φ t x ) φ t x p 5
6 for all P P n o, x R n \{o}, and all φ SL(n). Note that this together with (3) implies the SL(n) contravariance of degree p of µ d provided that µ is SL(n) contravariant of degree p. Surface area measures and their L p analogs were already defined in the introduction. It will be convenient for us to write S p (P, x) = S p (P, x ) x p for x R n \{o}. Similar conventions will apply to the measures S (P, ) and S o (P, ) := S(P, ) S (P, ). The following lemma guarantees that surface area measures and their L p analogs are GL(n) contravariant valuations. Lemma 5. The measures S, S, S o and the L p surface area measures S p for p are GL(n) contravariant valuations of degree p on P n o. Proof. First, assume that p. Given a phomogeneous function f : R n \{o} R, define for v S n, { f 0 t 0, (t, v) = t p f(v) t > 0. It was shown in [33] that for such functions f : R S n R the expression Z f P = V n (F (P, v))f (h(p, v), v) v N (P ) is a realvalued valuation. Here, F (P, v) denotes the facet of P with outer unit normal vector v. Let ω be a Borel set on S n and take f(x) = I ω (x/ x ) x p, where I ω denotes the indicator function of ω. Then we have S p (P, ω) = Z f P and we see that S p is a measure valued valuation. For φ GL(n) and v S n we clearly have as well as v N (P ) φ t v N (φp ) V n (F (φp, φ t v )) = φ t v det φ V n (F (P, v)). The transformation behavior of the support function with respect to the general linear group and the homogeneity of f yield f (h(φp, φ t v ), φ t v ) = f (h(p, φ t φ t v ), φ t v ) Therefore we obtain that = f ( φ t v h(p, v), φ t v φ t v) = φ t v f (h(p, v), φ t v). V n (F (φp, φ t v ))f (h(φp, φ t v ), φ t v ) = det φ V n (F (P, v))f (h(p, v), φ t v). 6
7 This immediately implies the GL(n) contravariance of degree p of S p. For p =, the same proof yields the desired properties for S. By changing the definition of f to { f 0 t < 0, (t, v) = f(v) t 0 we obtain these properties for the surface area measure S. Therefore we also have them for S o = S S. For p > 0, the L p cosine transform of a signed finite Borel measure µ on S n is defined by C p µ(u) = u v p dµ(v), u S n. S n We need the following injectivity result: For a finite signed Borel measure µ on S n (5) C µ = 0 = µ(ω) = µ( ω), for each Borel set ω on S n. In the proof of our classification results for positive p we will make use of known characterizations of function valued valuations. Therefore, we need a translation of SL(n) contravariance to such valuations. For positive p, a function Z : Po n C p + (R n ) is called SL(n) contravariant if Z(φP )(x) = Z P (φ x) for all P Po n, each φ SL(n), and all x R n. The next two results were established in [3] and [33], respectively. Theorem 6. If Z : P n o C + (Rn ), + is an even SL(n) contravariant valuation, then there exist constants c, c 2 R such that for every P P n o. Z P = C (c S(P, ) + c 2 S (P, )) Theorem 7. Let p R + \{}. If Z : P n o C + p (R n ), + is an even SL(n) contravariant valuation, then there exists a constant c R such that for every P P n o. Z P = C p (cs p (P, )) As announced in the introduction, we briefly describe where the notion of G contravariance comes from. The basis of the L p BrunnMinkowski theory is the following addition for convex bodies. Let p and suppose that P, Q Po n contain the origin in their interiors. For ε > 0 there exists a unique convex body P + p ε Q such that h(p + p ε Q, ) p = h(p, ) p + εh(q, ) p. 7
8 Using this addition, Lutwak [26] extended the classical case p = in order to prove that for the volume V and all p, V (P + p ε Q) V (P ) lim = h(q, u) p ds p (P, u). ε 0 + ε p S n This limit is called the L p mixed volume of P and Q and is an important notion of the L p BrunnMinkowski theory. The L p mixed volume is denoted by V p (P, Q). Clearly we have V p (φp, Q) = det φ V p (P, φ Q) for all φ GL(n). The above integral representation and the fact that h(φ Q, u) = h(q, φ t u) imply that h(q, u) p ds p (φp, u) = det φ h(q, φ t u) p ds p (P, u). S n S n Note that pth powers of support functions are homogeneous of degree p. Moreover, Kiderlen [9] showed that differences of pth powers of support functions are dense in the space of continuous functions on S n. So the last equation immediately implies that L p surface area measures are GL(n) contravariant of degree p. 3 Functional equations Let λ (0, ) and p R. We define two families of linear maps on R n by and φ λ e = e, φ λ e 2 = ( λ)e + λe 2, φ λ e k = e k for 3 k n, ψ λ e = ( λ)e + λe 2, ψ λ e 2 = e 2, ψ λ e k = e k for 3 k n. The following functional equation for f : R + R n \{o} R will play a key role: ) ( ) (6) f(s, x) = λ p n f (sλ n, φ t λ x + ( λ) p n f s( λ) n, ψ t λ x. The next result proves that a function which satisfies (6) at certain points is homogeneous in its first argument. Lemma 8. Let p R and suppose that a function f : R + R n \{o} R satisfies (6). If x R n \{o} is a fixed point of φ t λ and ψt λ and f(, x) is bounded from below on some open interval then f(s, x) = s n p f(, x) for every s > 0. 8
9 Proof. From (6) we see that (7) f(s n, x) = λ p n f(s n λ n, x) + ( λ) p n f(s n ( λ) n, x) for every s > 0 and λ (0, ). Define a function g : R + R by g(s) = f(s n, x). Then, for every s > 0 and λ (0, ), equation (7) reads as (8) g(s) = λ p n g(sλ) + ( λ) p n g(s( λ)). Let a and b be arbitrary positive real numbers. Set s = a + b and λ = a(a + b). If we insert these particular values of s and λ in (8), then we have for all a, b > 0, (a + b) p n g(a + b) = a p n g(a) + b p n g(b). Thus the function t t p/n g(t) solves Cauchy s functional equation on R + and, by assumption, it is bounded from below on some open interval. It is well known (see e.g. [, Corollary 9]) that this implies that t p n g(t) = tg() and hence g(t) = t p n g(). Finally, the definition of g immediately yields f(s, x) = g(s n ) = s n p g() = s n p f(, x). Next, we are going to show that special solutions of (6) are determined by their values on a small set. Lemma 9. Let p R and suppose that f : R + R n \{o} R has the following properties: (i) f satisfies (6). (ii) f is positively homogeneous of degree p in the second argument. (iii) For every s R + the function f(s, ) has countable support if restricted to S n. (iv) For every x R n \{o} the function f(, x) is bounded from below on some open interval. 9
10 (v) For each π SL(n) which is induced by a permutation matrix and all (s, x) R + R n \{o} (9) f(s, πx) = f(s, x). If then f(s, x) = 0 for every (s, x) R + {±e }, f(s, x) = 0 for every (s, x) R + R n \ lin{e + + e n }. Proof. Note that (6) gives for all (s, x) R + R n \{o} and each λ (0, ), (0) f(s, ψ t λ x) = λ p n f(sλ n, φ t λ ψ t λ x) + ( λ) p n f(s( λ) n, x), () f(s, φ t λ x) = λ p p n f(sλ n, x) + ( λ) n f(s( λ) n, ψ t λ φ t λ x). For j n we use induction on the number j of nonvanishing coordinates to prove that f(s, x) = 0 for all s > 0 and every x with j nonvanishing coordinates. Let j =. Since we can always find a permutation matrix π SL(n) with e i = πe, we have f(s, ±e i ) = f(s, ±πe ) = f(s, ±e ) = 0. Since f is positively homogeneous in the second argument we infer that f(s, x) = 0 for every x with one nonvanishing coordinate. Let j < n and suppose that f(s, x) = 0 for every s > 0 and every x o with at most j nonvanishing coordinates. By (9) it is enough to prove f(s, x) = 0 for x = x e + +x j+ e j+ with x,..., x j+ 0. Suppose that 0 < x < x 2 or x 2 < x < 0 and set λ = x /x 2. Then ψ t λ x = x 2e 2 + x 3 e x j+ e j+, φ t λ ψ t λ x = x e 2 + x 3 e x j+ e j+. Relation (0) and the induction hypothesis show f(s( λ) n, x) = 0 for every s R +, and hence f(s, x) = 0 for every s R +. If 0 < x 2 < x or x < x 2 < 0, then set λ = (x x 2 )/x. Thus φ t λ x = x e + x 3 e x j+ e j+, ψλ t φ t λ x = x 2e + x 3 e x j+ e j+. Relation () and the induction hypothesis show f(sλ n, x) = 0 for every s R +, and hence f(s, x) = 0 for every s R +. 0
11 If sgn(x ) sgn(x 2 ), set λ = x /(x x 2 ). Then φ t λ x = x e + x 3 e x j+ e j+, ψ t λ x = x 2e 2 + x 3 e x j+ e j+. It follows directly from (6) that f(s, x) = 0 for every s R +. In conclusion, we proved that for x = x e + +x j+ e j+ with x,..., x j+ 0 and x x 2 we have f(s, x) = 0 for every s R +. This and (9) actually show that f(s, x) = 0 for every s R + and each x = x e + + x j+ e j+ where at least two coordinates are different. It remains to prove that f(s, x) = 0 for every s R + and x = x e + + x e j+ with x 0. By the homogeneity of f in its second argument it suffices to prove that f(s, x) = 0 for every s R + and x = e + + e j+ or x = e e j+. We consider only the case x = e + + e j+ ; the other one is treated similarly. Let 0 < λ < and set y = λe + e e j+2. Note that ψ t λ y = e 2 + e e j+2, φ t λ ψ t λ y = λe 2 + e e j+2. By what we have already shown, relation (9), and (0) we arrive at f(s, ψ t λ y) = ( λ) p n f(s( λ) n, y). From (9) and Lemma 8 we infer that f(, ψ t λ y) is positively homogeneous of degree n p, and hence ( λ) f(, ψ t λ y) = f(, y) for all 0 < λ <. If f(, ψ t λ y) = f(, e 2 + e e j+2 ) were nonzero, then f(, y) = f(, λe + e e j+2 ) would therefore be nonzero for all 0 < λ <. But since f is positively homogeneous in the second argument, this would contradict the assumption that f(, ) has countable support on S n. By homogeneity, for each s R +, we have f(s, ψ t λ y) = 0 and hence (9) gives f(s, e + + e j+ ) = 0. This concludes the induction. We showed that f(s, x) = 0 for s R + and points x with at most n nonvanishing coordinates. As in the first part of the induction we see that also f(s, x) = 0 for s R + and points x with n nonvanishing coordinates provided that at least two of them are different. 4 The case p = Lemma 0. Suppose that µ : P n o M d (S n ) is SL(n) contravariant of degree. Then there exists a constant a R + such that µ(st, ) = as n (δ e + δ e )
12 for every s R + and µ({o}, ) = 0, where T denotes the (n )dimensional standard simplex in e. Proof. First, we show that µ(p, ) is supported at ±e k provided that P e k. So let P e k and set f(p, x) = µ(p, x ) x for x R n \{o}. Suppose that x R n is given with x j 0 for some j k. For t R define φ SL(n) by Since φp = P we get from (4) that φe k = e k + te j, φe i = e i, i {,..., n}\{k}. f(p, x) = f(φp, x) = f(p, φ t x) = f(p, (x,..., x k, x k + tx j, x k+,..., x n )). If f(p, x) 0, then f(p, (x,..., x k, x k + tx j, x k+,..., x n )) 0 for all t R. Since these correspond to uncountably many points on S n and f(p, ) restricted to S n has countable support, f(p, x) has to be zero. Consequently, f(p, ) is supported only at ±e k. In particular, the measure µ({o}, ) has to be supported at ±e as well as ±e 2. Therefore, it has to be zero. It remains to prove the formula for µ(st, ). We already know that f(st, ) is supported only at ±e. Thus Define ψ SL(n) by f(st, x) = a (s)δ e ( x ) x + a 2 (s)δ e ( x ) x. ψe = e, ψe 2 = e 3, ψe 3 = e 2, ψe i = e i, 4 i n. Then by (4) and the relation st = ψ(st ) we have f(st, e ) = f(st, e ) and therefore a (s) = a 2 (s). Finally, define τ SL(n) by τe = s n e, τe i = se i, 2 i n. Then st = τt and relation (4) again show that f(st, e ) = s n f(t, e ), which proves a (s) = s n a(). Now set a = a(). Now, we are in a position to prove Theorem. For the reader s convenience we will repeat its statement. 2
13 Theorem. A map µ : P n o M(S n ) is an SL(n) contravariant valuation of degree if and only if there exist constants c, c 2, c 3, c 4 R with c, c 2 0 and c + c 3 0, c 2 + c 4 0 such that (2) µ(p, ) = c S(P, ) + c 2 S( P, ) + c 3 S (P, ) + c 4 S ( P, ) for every P P n o. Proof. Set f(p, x) = µ d (P, x ) x, x R n \{o}, and define constants By Lemma 0 we have Hence, the function d = (n )!(f(t, e ) f(t n, e )), d 2 = (n )!f(t n, e ). f(st, x) = f(t, e )s n (δ e ( x ) + δ e ( x )) x. g(p, x) = f(p, x) d S o (P, x) d 2 S o ( P, x) is a valuation which vanishes on st. For λ (0, ) let H λ be the hyperplane through o with normal vector λe ( λ)e 2. Note that (st n ) H + λ = sφ λt n, (st n ) H λ = sψ λt n and (st n ) H λ = sφ λ T. So the valuation property of g implies that for all (s, x) R + R n \{o}, g(st n, x) + g(sφ λ T, x) = g(sφ λ T n, x) + g(sψ λ T n, x). By (4) we have g(sφ λ T, x) = λ /n g(sλ /n T, φ t λ x), g(sφ λt n, x) = λ /n g(sλ /n T n, φ t λ x), and g(sψ λ T n, x) = ( λ) /n g(s( λ) /n T n, ψλ t x). This and the fact that g vanishes on multiples of T proves (3) g(st n, x) = λ n g(sλ n T n, φ t λ x) + ( λ) n g(s( λ) n T n, ψ t λ x). Consequently, the map (s, x) g(st n, x) satisfies (6). By Lemma 8 we know that g(st n, e 3 ) = s n g(t n, e 3 ) for all positive s. But g(st n, e ) = g(st n, e 3 ) and g(t n, e ) = 0 by definition. Thus g(st n, e ) = 0 for all s R +. For x = x e + x 2 e 2 with x > 0 and x 2 < 0 set λ = x x x 2. Evaluating (3) for this x and λ, Lemma 8 together with the homogeneity of g of degree in the second argument, the equality g(st n, e ) = 0 and the definition of λ show that g(st n, x) = λ n g(sλ n T n, x e ) + ( λ) n g(s( λ) n T n, x 2 e 2 ) = λ g(st n, e ) + λ g(st n, e 2 ) x x 2 = g(st n, e 2 ). x x 2 3
14 Since g(st n, e 2 ) = g(st n, e ) and g(st n, ) has at most countable support if restricted to S n, we obtain from the last lines that also g(st n, e ) = 0. From Lemma 9 we further deduce that g(st n, ) is supported only at ± e + +e n. Define constants d 3 = g(t n, e + + e n ) S (T n, e + + e n ), d 4 = g(t n, e + + e n ) S ( T n, e + + e n ). Thus, by Lemma 8, we have g(st n, x) = d 3 S (st n, x) + d 4 S ( st n, x) and consequently µ d (st n, x) = d S(sT n, x)+d 2 S( st n, x)+(d 3 d )S (st n, x)+(d 4 d 2 )S ( st n, x). By Lemma 4 this proves (2) for the discrete part µ d. Note that P C µ(p, ) is a function from Po n to C + (Rn ) which is an even SL(n) contravariant valuation. The linearity of the cosine transform gives C µ(p, ) = C µ c (P, ) + C µ d (P, ). From Theorem 6 we know that C µ(p, ) = C (d 5 S(P, ) + d 6 S (P, )). By the discrete case we just established we know that and hence also C µ d (P, ) = C (d 7 S(P, ) + d 8 S (P, )), C µ c (P, ) = C (d 9 S(P, ) + d 0 S (P, )). In particular, we have C µ c (st, ) = d 9 C S(sT, ). By the injectivity property (5) we know that µ c (st, e ) + µ c (st, e ) = 2s n d 9 /(n )!. But since µ c (st, ) is continuous we have d 9 = 0. Thus C µ c (st n, )(e i e j ) = d 0 C S (st n, )(e i e j ) = 0 for i j n. So µ c (st n, ) has to be concentrated on each hyperplane {x i = x j }, and consequently it is concentrated at the two points ± e + + e n. The continuity of µ c (st n, ) therefore implies µ c (st n, ) = 0. So by the discrete case we have µ(st n, ) = µ d (st n, ) = c S(sT n, ) + c 2 S( st n, ) + c 3 S (st n, ) + c 4 S ( st n, ). for some constants c,..., c 4 R. Next, we want to prove that µ({o}, ) = 0. For s > 0 define φ SL(n) by φe = s e, φe 2 = se 2, φe k = e k, 3 k n. 4
15 Since φ{o} = {o} the SL(n) contravariance of µ implies µ({o}, S n ) = x dµ(φ{o}, x) S n = (s 2 x 2 + s 2 x x x 2 n) 2 dµ({o}, x). S n Take the limit s in the above equation. Then Fatou s lemma implies µ({o}, S n ) µ({o}, {x 0}). Since µ({o}, ) is finite, µ({o}, ) is supported at {x = 0}. Similarly, one shows that µ({o}, ) is supported at {x j = 0} for j = 2,..., n. This immediately implies that µ({o}, ) = 0. Lemma 4 and the fact that S(P, ) = S o (P, ) + S (P, ) therefore prove µ(p, ) = c S(P, ) + c 2 S( P, ) + c 3 S (P, ) + c 4 S ( P, ) = c S o (P, ) + c 2 S o ( P, ) + (c + c 3 )S (P, ) + (c 2 + c 4 )S ( P, ) for every P P n o. In particular as well as 0 µ(t n, e ) = c 2 /(n )! and 0 µ(t n, e ) = c /(n )!, 0 µ(t n, e + + e n ) = n(c + c 3 )/(n )!, 0 µ(t n, e + + e n ) = n(c 2 + c 4 )/(n )!, which proves the asserted relations for the constants. 5 The case p We begin by studying SL(n) contravariant valuations on lower dimensional polytopes. Lemma 2. Suppose that p and µ : Po n M(S n ) is an SL(n) contravariant valuation of degree p which vanishes on polytopes of dimension less than n. Let s R +. If µ(st, ) is supported only at the two points ±e, then µ(st, ) = 0. Proof. For λ (0, ) let H λ denote the hyperplane containing the origin with normal vector λe 2 ( λ)e 3. The valuation property of µ yields (4) µ(st, ) + µ(st H λ, ) = µ(st H + λ, ) + µ(st H λ, ). 5
16 Define maps σ, τ SL(n) by and Since σe = λ e, σe 2 = e 2, σe 3 = ( λ)e 2 + λe 3, σe k = e k, 4 k n, τe = λ e, τe 2 = ( λ)e 2 + λe 3, τe k = e k, 3 k n. st H + λ = σ(st ) and st H λ = τ(st ) and since µ vanishes on st H λ, we have Thus by (4) we obtain µ(st, ) = µ(σ(st ), ) + µ(τ(st ), ). µ(st, ±e ) = µ(σ(st ), ±e ) + µ(τ(st ), ±e ) Since p, we get µ(st, ±e ) = The discrete case = µ(st, ± σ t e ) σ t e p + µ(st, ± τ t e ) τ t e p = (λ p + ( λ) p )µ(st, ±e ). The next result concerns simplicity for valuations with discrete images. Lemma 3. Let p and suppose that µ : Po n M d (S n ) is SL(n) contravariant of degree p. Then µ(st, ) = 0 for every s R + and µ({o}, ) = 0. Proof. By Lemma 2 it is enough to show that µ(p, ) is supported at ±e k provided that P e k. So let P e k and set f(p, x) = µ(p, x ) x p for x R n \{o}. Suppose that x R n is given with x j 0 for some j k. For t R define φ SL(n) by Since φp = P, we get from (4) that φe k = e k + te j, φe i = e i, i {,..., n}\{k}. f(p, x) = f(φp, x) = f(p, φ t x) = f(p, (x,..., x k, x k + tx j, x k+,..., x n )). If f(p, x) 0, then f(p, (x,..., x k, x k + tx j, x k+,..., x n )) 0 for all t R. Since these correspond to uncountably many points on S n and f(p, ) restricted to S n has countable support, f(p, x) has to be zero. Consequently, f(p, ) is supported only at ±e k. 6
17 The main result of this section is the following classification of SL(n) contravariant valuations of degree p with discrete images. Theorem 4. Let p R\{}. A map µ : P n o M d (S n ) is an SL(n) contravariant valuation of degree p if and only if there exist nonnegative constants c, c 2 R such that µ(p, ) = c S p (P, ) + c 2 S p ( P, ) for every P P n o. By Lemmas 4 and 3, this theorem will be an immediate consequence of the following Lemma 5. Let p and µ : Po n M(S n ) be an SL(n) contravariant valuation of degree p. If, for all s > 0, µ(st, ) = 0 and µ(st n, ) is discrete, then there exist constants c, c 2 0 such that for all s > 0 µ(st n, ) = c S p (st n, ) + c 2 S p ( st n, ). Proof. For λ (0, ) let H λ be the hyperplane through o with normal vector λe ( λ)e 2. Note that (st n ) H + λ = sφ λt n, (st n ) H λ = sψ λt n, and (st n ) H λ = sφ λ T. Set f(p, x) = µ(p, x ) x p. (s, x) R + R n \{o}, Then the valuation property of µ implies that for all f(st n, x) + f(sφ λ T, x) = f(sφ λ T n, x) + f(sψ λ T n, x). From (4) and the assumption that µ vanishes on multiples of T we deduce that f(st n, x) = λ p n f(sλ n T n, φ t λ x) + ( λ) p n f(s( λ) n T n, ψ t λ x). Consequently, the map (s, x) f(st n, x) satisfies (6). Evaluating this equality at the first canonical basis vector e proves (5) f(s, e ) = λ p n f(sλ n, e + ( λ)e 2 ) + ( λ) p n f(s( λ) n, ( λ)e ). But e 3 is a fixpoint of φ t λ and ψt λ and f(s, e 3) = f(s, e ) for all positive s, hence by Lemma 8 we get f(s, e ) = s n p f(, e ) for all s > 0. Thus we obtain from (5) that ( ( λ) p )f(s, e ) = λ p n f(sλ n, e + ( λ)e 2 ). 7
18 This shows in particular that f(sλ n, e + ( λ)e 2 ) = s n p f(λ n, e + ( λ)e 2 ) for all s > 0 and each λ (0, ). Consequently and hence f(sλ n, e + ( λ)e 2 ) = s n p λ p n f(, e + ( λ)e 2 ) ( ( λ) p )λ f(s, e ) = f(s, e + ( λ)e 2 ). Since p and for fixed s the function f(s, ) has countable support if restricted to S n, we see that f(s, e ) = 0 for all s > 0. Similarly, by looking at e instead of e in the above argument, we infer f(s, e ) = 0 for all s > 0. From Lemma 9 we deduce that f(s, ) is supported only at ± e + + e n. Define constants c = f(, e + + e n ) S p (T n, e + + e n ) and c 2 = f(, e + + e n ) S p ( T n, e + + e n ). Since ± e + + e n are fixpoints of φ t λ and ψt λ, we therefore have by Lemma 8 that f(s, x) = c S p (st n, x) + c 2 S p ( st n, x) for all s > 0 and all x S n. Now, relation (2) concludes the proof. 5.2 The case p > 0 Lemma 6. Let p R + \{}. If µ : P n o M(S n ) is an SL(n) contravariant valuation of degree p, then µ(st, ) = 0 for every s > 0 and µ({o}, ) = 0. Proof. By Lemma 2 it is enough to show that µ(p, ) is supported at ±e k provided that P e k. So let P e k. For t R and j k define φ SL(n) by φe k = e k + te j, φe i = e i, i {,..., n}\{k}. Note that φp = P. Therefore the SL(n) contravariance of µ implies µ(p, S n ) = µ(φp, S n ) = x p dµ(φp, x) S n ( = x x 2 k + (x k tx j ) 2 + x 2 k+ + + x2 n S n ) p 2 dµ(p, x) If µ(p, {x j 0}) 0, then Fatou s lemma shows that the last integral goes to infinity as t tends to infinity. But this would contradict the finiteness of µ(p, ). Thus µ(p, {x j 0}) = 0, which immediately implies that µ(p, ) is supported only at ±e k. 8
19 Now, we establish the characterization of L p surface area measures for positive p which was already announced in the introduction. We emphasize again that for positive p no homogeneity assumptions are needed. It suffices to assume SL(n) contravariance instead of GL(n) contravariance. Theorem 7. Let p R + \{}. A map µ : P n o M(S n ) is an SL(n) contravariant valuation of degree p if and only if there exist nonnegative constants c, c 2 R such that µ(p, ) = c S p (P, ) + c 2 S p ( P, ) for every P P n o. Proof. Note that C p µ(p, ) : Po n Thus, by Theorem 7, we have C + p (R n ) is an even SL(n) contravariant valuation. C p µ(st n, )(e i e j ) = c C p S p (st n, )(e i e j ) = 0. for i j n. So µ(st n, ) has to be concentrated on each hyperplane {x i = x j }, and consequently it is concentrated at the two points ± e + + e n. By Lemmas 5 and 6 there exist nonnegative constants c and c 2 such that µ(st n, x) = c S p (st n, x) + c 2 S p ( st n, x) for all s > 0 and x S n. Thus, Lemma 4 concludes the proof. 5.3 The case p 0 Next, we prove the GL(n) contravariant case for nonpositive p. We start with the simplicity in this case. Lemma 8. Let p 0 and µ : Po n M(S n ) be a GL(n) contravariant map of degree p. Then µ is simple, i.e. it vanishes on polytopes of dimension less than n. Proof. Without loss of generality we assume that P e. For s > 0 define φ GL(n) by φe = se, φe k = e k, 2 k n. Since φp = P, the GL(n) contravariance of µ yields µ(p, S n ) = µ(φp, S n ) = x p dµ(φp, x) S n ( ) x 2 p = s S n s 2 + x x 2 2 n µ(p, x). 9
20 Note that the integrand of the last integral is equal to ( + x 2 (/s2 ) ) p/2. Clearly, this function is greater than or equal to ( + /s 2 ) p/2 and hence µ(p, S n ) µ(p, S n )s ( + s ) p 2. 2 Now take the limit s and use the fact that µ(p, ) is a finite measure in order to arrive at µ(p, S n ) = 0. Hence µ(p, ) = 0. We are now in a position to prove Theorem 3 for nonpositive p. Lemma 9. Let p 0. A map µ : P n o M(S n ) is a GL(n) contravariant valuation of degree p if and only if there exist nonnegative constants c, c 2 R such that for every P P n o. µ(p, ) = c S p (P, ) + c 2 S p ( P, ) Proof. For λ (0, ) let H λ be the hyperplane through o with normal vector λe ( λ)e 2. Recall that T n H + λ = φ λt n, T n H λ = ψ λt n and T n H λ = φ λ T. The valuation property, the simplicity derived in Lemma 8, and the GL(n) contravariance of µ yield f dµ(t n, ) = λ S n f φ t λ dµ(t n, ) + ( λ) S n f ψ t λ dµ(t n, ). S n Thus, for every continuous phomogeneous f : R n \{o} R, we have (6) 0 = f φ t λ dµ(t n t ( λ)(f ψλ, ) + ) f dµ(t n, ). S n S n λ First, assume p < 0. Define f p (x) = ( (x x 2 ) 2 + x x2 n) p/2 for x R n \{o}. Note that f p is strictly positive and phomogeneous. Set g p (λ, x) = f p ψ t λ = An elementary calculation shows that ( (x ) x x x 2 n λ ) p 2. (7) λ ( λ)g p(λ, x) = g p (λ, x) + p g p (λ, x) 2 (x x p 2 ) 2 ( λ) 2, 20
21 and hence ( λ)(f p ψ t λ lim ) f p (8) λ 0 + λ = λ ( λ)g p(λ, x) λ=0 = f p (x) + pf p (x) 2 p (x x 2 ) 2. Clearly we have lim λ 0 +(f p φ t λ )(x) = lim λ 0 + ( (x x 2 λ = I {x =x 2 }(x)f p (x). ) 2 ( ) x2 x x + x x 2 n λ If we take the limit λ 0 + in (6) and are allowed to interchange limit and integrals, then ( 0 = f p (x) I {x =x 2 }(x) + pf p (x) 2 p (x x 2 ) 2) dµ(t n, x). S n Since the last integrand is less than or equal to zero and equal to zero precisely for points x with x = x 2, we deduce that µ(t n, ) is concentrated on {x = x 2 }. Since by the GL(n) contravariance µ(t n, ) is invariant with respect to coordinate changes we get that µ(t n, ) is concentrated at the points ± e + + e n. In particular, µ(t n, ) is discrete. We still have to prove that we are actually allowed to interchange limit and integrals in the above argument. Since g p (λ, x) is obviously bounded on (0, /2] S n, equations (7) and (8), the mean value theorem, and the dominated convergence theorem show that we can interchange limit and integration in the second integral. In order to apply the dominated convergence theorem for the first integral it is enough to prove that the function f p φ t λ is bounded from above on (0, ) Sn. Recall that ( (x ) f p φ t λ (x) = x 2 ( 2 + x + x ) ) p 2 x x x 2 n. λ λ Define a set U = {x S n : (x x 2 ) 2 /25 or x x 2 n /25}. By the negativity of p, the function f p φ t λ is obviously bounded on (0, ) U. So suppose that x S n \U, i.e. (x x 2 ) 2 < /25 and x x2 n < /25. From these two inequalities and the fact that (x x 2 ) 2 + 2x x 2 + x x2 n = we get 2x x 2 > 2/25 = 23/25. Since x 2 we further obtain (9) x > ) p 2
22 Assume that there exists a λ (0, ) such that ( x x 2 Then we obviously must have λ ) 2 ( + x + x ) 2 x 2 + x x 2 n < λ 25. (20) as well as (2) ( x x 2 λ ) 2 < 25 ( x + x ) 2 x 2 < λ 25. But by the reverse triangle inequality, (9), and (20) we obtain x + x 2 x λ x x 2 x λ > 5, which contradicts (2). Consequently, f p φ t λ is bounded also on (0, ) S n \U, which immediately gives the desired result. For the case p = 0 set h(x) = x 3 f (x), x R n \{o}. Note that h is a nonnegative continuous 0homogeneous function. Thus 0 = = h φ t λ dµ(t n t ( λ)(h ψλ, ) + ) h dµ(t n, ) S n S n λ x 3 (f φ t λ ) dµ(t n, ) + x 3 ( λ)(f t ψλ ) f dµ(t n, ) S n S n λ As before, we can take the limit λ 0 + in this equation and interchange limit and integration in order to arrive at 0 = x 3 f (x) ( I {x =x 2 }(x) f (x) 2 (x x 2 ) 2) dµ(t n, ). S n Thus µ(t n, ) has to be concentrated on {x 3 = 0} {x = x 2 }. Since µ(t n, ) is invariant under coordinate changes, µ(t n, ) has to be supported at { {xi = 0} {x j = x k } : i, j, k n with distinct i, j, k }. Suppose that x S n is a point in this intersection. Then either all coordinates of x are equal, or at least two are different, say x and x 2. But then all other coordinates 22
23 have to be zero. At least one of x and x 2 has to be nonzero, say x. Since x {x = 0} {x 2 = x 3 }, we have x 2 = 0 and thus x = ±. This implies that µ(t n, ) is concentrated at the points ±e i and ± e + + e n. So also in this case µ(t n, ) is discrete. By Lemmas 5 and 8 we infer that there exist constants c, c 2 0 such that for all s > 0 µ(st n, ) = c S p (st n, ) + c 2 S p ( st n, ). By the simplicity of µ we also have µ({o}, ) = 0. Lemma 4 therefore concludes the proof. Acknowledgments The work of the authors was supported by Austrian Science Fund (FWF) Project P23639N8. References [] J. Aczél and J. Dhombres. Functional equations in several variables, volume 3 of Encyclopedia of Mathematics and its Applications. Cambridge University Press, Cambridge, 989. With applications to mathematics, information theory and to the natural and social sciences. [2] S. Alesker. Continuous rotation invariant valuations on convex sets. Ann. of Math. (2) 49 (999), [3] S. Alesker. Description of translation invariant valuations on convex sets with solution of P. McMullen s conjecture. Geom. Funct. Anal. (200), [4] S. Alesker, A. Bernig and F. Schuster. Harmonic analysis of translation invariant valuations. Geom. Funct. Anal. 2 (20), [5] A. Bernig. A Hadwigertype theorem for the special unitary group. Geom. Funct. Anal. 9 (2009), [6] A. Bernig and J. H. G. Fu. Hermitian integral geometry. Ann. of Math (2) 73 (20), [7] K.S. Chou and X.J. Wang. The L p Minkowski problem and the Minkowski problem in centroaffine geometry. Adv. Math. 205 (2006),
24 [8] A. Cianchi, E. Lutwak, D. Yang and G. Zhang. Affine MoserTrudinger and MorreySobolev inequalities. Calc. Var. Partial Differential Equations 36 (2009), [9] J. Fu. Structure of the unitary valuation algebra. J. Differential Geom. 72 (2006)(3), [0] R. Gardner and A. Giannopoulos. pcrosssection bodies. Indiana Univ. Math. J. 48 (999), [] R. J. Gardner. Geometric tomography, volume 58 of Encyclopedia of Mathematics and its Applications. Cambridge University Press, New York, second edition, [2] P. Gruber. Convex and Discrete Geometry. Springer, Berlin, [3] C. Haberl. Minkowski valuations intertwining the special linear group. J. Eur. Math. Soc., in press. [4] C. Haberl. Blaschke valuations. Amer. J. Math. 33 (20), [5] C. Haberl and M. Ludwig. A characterization of L p intersection bodies. Int. Math. Res. Not. (2006), Article ID 0548, 29 pages. [6] C. Haberl, E. Lutwak, D. Yang and G. Zhang. The even Orlicz Minkowski problem. Adv. Math 224 (200), [7] C. Haberl and F. Schuster. Asymmetric affine L p Sobolev inequalities. J. Funct. Anal. 257 (2009), [8] C. Haberl and F. Schuster. General L p affine isoperimetric inequalities. J. Differential Geom. 83 (2009), 26. [9] M. Kiderlen. Stability results for convex bodies in geometric tomography. Indiana Univ. Math. J. 57 (2008), [20] D. A. Klain. Invariant valuations on starshaped sets. Adv. Math. 25 (997), [2] M. Ludwig. Ellipsoids and matrix valued valuations. Duke Math. J. 9 (2003), [22] M. Ludwig. Minkowski valuations. Trans. Amer. Math. Soc. 357 (2005), [23] M. Ludwig. Minkowski areas and valuations. J. Differential. Geom. 86 (200),
25 [24] M. Ludwig. Fisher information and matrix valued valuations. Adv. Math. 226 (20), [25] M. Ludwig and M. Reitzner. A classification of SL(n) invariant valuations. Ann. Math. (2) 72 (200), [26] E. Lutwak. The BrunnMinkowskiFirey theory. I. Mixed volumes and the Minkowski problem. J. Differential Geom. 38 (993), [27] E. Lutwak. The BrunnMinkowskiFirey theory. II: Affine and geominimal surface areas. Adv. Math. 8 (996), [28] E. Lutwak, D. Yang and G. Zhang. L p affine isoperimetric inequalities. J. Differential Geom. 56 (2000), 32. [29] E. Lutwak, D. Yang and G. Zhang. A new ellipsoid associated with convex bodies. Duke Math. J. 04 (2000), [30] E. Lutwak, D. Yang and G. Zhang. Sharp affine L p Sobolev inequalities. J. Differential Geom. 62 (2002), [3] E. Lutwak, D. Yang and G. Zhang. Orlicz centroid bodies. J. Differential Geom. 84 (200), [32] E. Lutwak, D. Yang and G. Zhang. Orlicz projection bodies. Adv. Math. 223 (200), [33] L. Parapatits. SL(n)contravariant L p Minkowski valuations. Trans. Amer. Math. Soc., in press. [34] L. Parapatits and F. Schuster. The Steiner Formula for Minkowski Valuations. Adv. Math., in press. [35] A. V. Pogorelov. The Minkowski Multidimensional Problem. V. H. Winston & Sons, Washington, D. C., 978. [36] D. Ryabogin and A. Zvavitch. The Fourier transform and Firey projections of convex bodies. Indiana Univ. Math. J. 53 (2004)(3), [37] R. Schneider. Kinematische Berührmaße für konvexe Körper. Abh. Math. Sem. Univ. Hamburg 44 (975), [38] R. Schneider. Convex bodies: the Brunn Minkowski theory. Cambridge Univ. Press, Cambridge,
26 [39] F. Schuster. Valuations and BusemannPetty type problems. Adv. Math. 29 (2008), [40] F. Schuster. Crofton measures and Minkowski valuations. Duke Math. J. 54 (200), 30. [4] F. Schuster and T. Wannerer. GL(n) contravariant Minkwoski valuations. Trans. Amer. Math. Soc., in press. [42] C. Schütt and E. Werner. Surface bodies and paffine surface area. Adv. Math. 87 (2004), [43] A. Stancu. On the number of solutions to the discrete twodimensional L 0  Minkowski problem. Adv. Math. 80 (2003), [44] E. Werner and D. Ye. Inequalities for mixed paffine surface area. Math. Ann. 347 (200), [45] V. Yaskin and M. Yaskina. Centroid bodies and comparison of volumes. Indiana Univ. Math. J. 55 (2006), Christoph Haberl University of Salzburg Hellbrunnerstraße Salzburg, Austria Lukas Parapatits Vienna University of Technology Institute of Discrete Mathematics and Geometry Wiedner Hauptstraße 80/ Vienna, Austria 26
Intersection bodies and valuations
Intersection bodies and valuations Monika Ludwig Dedicated to Prof. Erwin Lutwak on the occasion of his sixtieth birthday Abstract All GL(n) covariant starbodyvalued valuations on convex polytopes are
More information(Basic definitions and properties; Separation theorems; Characterizations) 1.1 Definition, examples, inner description, algebraic properties
Lecture 1 Convex Sets (Basic definitions and properties; Separation theorems; Characterizations) 1.1 Definition, examples, inner description, algebraic properties 1.1.1 A convex set In the school geometry
More informationF. ABTAHI and M. ZARRIN. (Communicated by J. Goldstein)
Journal of Algerian Mathematical Society Vol. 1, pp. 1 6 1 CONCERNING THE l p CONJECTURE FOR DISCRETE SEMIGROUPS F. ABTAHI and M. ZARRIN (Communicated by J. Goldstein) Abstract. For 2 < p
More informationSOLUTIONS TO EXERCISES FOR. MATHEMATICS 205A Part 3. Spaces with special properties
SOLUTIONS TO EXERCISES FOR MATHEMATICS 205A Part 3 Fall 2008 III. Spaces with special properties III.1 : Compact spaces I Problems from Munkres, 26, pp. 170 172 3. Show that a finite union of compact subspaces
More informationSign changes of Hecke eigenvalues of Siegel cusp forms of degree 2
Sign changes of Hecke eigenvalues of Siegel cusp forms of degree 2 Ameya Pitale, Ralf Schmidt 2 Abstract Let µ(n), n > 0, be the sequence of Hecke eigenvalues of a cuspidal Siegel eigenform F of degree
More informationISOMETRIES OF R n KEITH CONRAD
ISOMETRIES OF R n KEITH CONRAD 1. Introduction An isometry of R n is a function h: R n R n that preserves the distance between vectors: h(v) h(w) = v w for all v and w in R n, where (x 1,..., x n ) = x
More informationMATHEMATICAL BACKGROUND
Chapter 1 MATHEMATICAL BACKGROUND This chapter discusses the mathematics that is necessary for the development of the theory of linear programming. We are particularly interested in the solutions of a
More informationBANACH AND HILBERT SPACE REVIEW
BANACH AND HILBET SPACE EVIEW CHISTOPHE HEIL These notes will briefly review some basic concepts related to the theory of Banach and Hilbert spaces. We are not trying to give a complete development, but
More informationTHEORY OF SIMPLEX METHOD
Chapter THEORY OF SIMPLEX METHOD Mathematical Programming Problems A mathematical programming problem is an optimization problem of finding the values of the unknown variables x, x,, x n that maximize
More informationNo: 10 04. Bilkent University. Monotonic Extension. Farhad Husseinov. Discussion Papers. Department of Economics
No: 10 04 Bilkent University Monotonic Extension Farhad Husseinov Discussion Papers Department of Economics The Discussion Papers of the Department of Economics are intended to make the initial results
More informationand s n (x) f(x) for all x and s.t. s n is measurable if f is. REAL ANALYSIS Measures. A (positive) measure on a measurable space
RAL ANALYSIS A survey of MA 641643, UAB 19992000 M. Griesemer Throughout these notes m denotes Lebesgue measure. 1. Abstract Integration σalgebras. A σalgebra in X is a nonempty collection of subsets
More information2.3 Convex Constrained Optimization Problems
42 CHAPTER 2. FUNDAMENTAL CONCEPTS IN CONVEX OPTIMIZATION Theorem 15 Let f : R n R and h : R R. Consider g(x) = h(f(x)) for all x R n. The function g is convex if either of the following two conditions
More informationContinued Fractions and the Euclidean Algorithm
Continued Fractions and the Euclidean Algorithm Lecture notes prepared for MATH 326, Spring 997 Department of Mathematics and Statistics University at Albany William F Hammond Table of Contents Introduction
More informationMathematics Course 111: Algebra I Part IV: Vector Spaces
Mathematics Course 111: Algebra I Part IV: Vector Spaces D. R. Wilkins Academic Year 19967 9 Vector Spaces A vector space over some field K is an algebraic structure consisting of a set V on which are
More informationFUNCTIONAL ANALYSIS LECTURE NOTES: QUOTIENT SPACES
FUNCTIONAL ANALYSIS LECTURE NOTES: QUOTIENT SPACES CHRISTOPHER HEIL 1. Cosets and the Quotient Space Any vector space is an abelian group under the operation of vector addition. So, if you are have studied
More informationarxiv: v1 [math.pr] 22 Aug 2008
arxiv:0808.3155v1 [math.pr] 22 Aug 2008 On independent sets in purely atomic probability spaces with geometric distribution Eugen J. Ionascu and Alin A. Stancu Department of Mathematics, Columbus State
More informationMA651 Topology. Lecture 6. Separation Axioms.
MA651 Topology. Lecture 6. Separation Axioms. This text is based on the following books: Fundamental concepts of topology by Peter O Neil Elements of Mathematics: General Topology by Nicolas Bourbaki Counterexamples
More informationAn Arithmetic Proof of John s Ellipsoid Theorem
An Arithmetic Proof of John s Ellipsoid Theorem Peter M. Gruber and Franz E. Schuster Abstract. Using an idea of Voronoi in the geometric theory of positive definite quadratic forms, we give a transparent
More informationOur goal first will be to define a product measure on A 1 A 2.
1. Tensor product of measures and Fubini theorem. Let (A j, Ω j, µ j ), j = 1, 2, be two measure spaces. Recall that the new σ algebra A 1 A 2 with the unit element is the σ algebra generated by the
More informationDuality of linear conic problems
Duality of linear conic problems Alexander Shapiro and Arkadi Nemirovski Abstract It is well known that the optimal values of a linear programming problem and its dual are equal to each other if at least
More informationINDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS
INDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS STEVEN P. LALLEY AND ANDREW NOBEL Abstract. It is shown that there are no consistent decision rules for the hypothesis testing problem
More informationNumerical Analysis Lecture Notes
Numerical Analysis Lecture Notes Peter J. Olver 5. Inner Products and Norms The norm of a vector is a measure of its size. Besides the familiar Euclidean norm based on the dot product, there are a number
More informationGeometrical Characterization of RNoperators between Locally Convex Vector Spaces
Geometrical Characterization of RNoperators between Locally Convex Vector Spaces OLEG REINOV St. Petersburg State University Dept. of Mathematics and Mechanics Universitetskii pr. 28, 198504 St, Petersburg
More informationIdeal Class Group and Units
Chapter 4 Ideal Class Group and Units We are now interested in understanding two aspects of ring of integers of number fields: how principal they are (that is, what is the proportion of principal ideals
More informationKOLMOGOROVSMIRNOV ISOMETRIES OF THE SPACE OF GENERALIZED DISTRIBUTION FUNCTIONS
KOLMOGOROVSMIRNOV ISOMETRIES OF THE SPACE OF GENERALIZED DISTRIBUTION FUNCTIONS LAJOS MOLNÁR* AND PATRĆIA SZOKOL** Abstract. In this paper we describe the structure of surjective isometries of the space
More informationLecture Notes on Measure Theory and Functional Analysis
Lecture Notes on Measure Theory and Functional Analysis P. Cannarsa & T. D Aprile Dipartimento di Matematica Università di Roma Tor Vergata cannarsa@mat.uniroma2.it daprile@mat.uniroma2.it aa 2006/07 Contents
More informationProperties of BMO functions whose reciprocals are also BMO
Properties of BMO functions whose reciprocals are also BMO R. L. Johnson and C. J. Neugebauer The main result says that a nonnegative BMOfunction w, whose reciprocal is also in BMO, belongs to p> A p,and
More information1. Introduction. PROPER HOLOMORPHIC MAPPINGS BETWEEN RIGID POLYNOMIAL DOMAINS IN C n+1
Publ. Mat. 45 (2001), 69 77 PROPER HOLOMORPHIC MAPPINGS BETWEEN RIGID POLYNOMIAL DOMAINS IN C n+1 Bernard Coupet and Nabil Ourimi Abstract We describe the branch locus of proper holomorphic mappings between
More informationSeparation Properties for Locally Convex Cones
Journal of Convex Analysis Volume 9 (2002), No. 1, 301 307 Separation Properties for Locally Convex Cones Walter Roth Department of Mathematics, Universiti Brunei Darussalam, Gadong BE1410, Brunei Darussalam
More informationCOUNTEREXAMPLES TO CONVEXITY OF kintersection BODIES.
COUNTEREXAPES TO CONVEXITY OF kintersection BODIES. VADYSAV YASIN Abstract. It is a wellknown result due to Busemann that the intersection body of an originsymmetric convex body is also convex. oldobsky
More informationCHAPTER SIX IRREDUCIBILITY AND FACTORIZATION 1. BASIC DIVISIBILITY THEORY
January 10, 2010 CHAPTER SIX IRREDUCIBILITY AND FACTORIZATION 1. BASIC DIVISIBILITY THEORY The set of polynomials over a field F is a ring, whose structure shares with the ring of integers many characteristics.
More informationThe fundamental group of the Hawaiian earring is not free (International Journal of Algebra and Computation Vol. 2, No. 1 (1992), 33 37) Bart de Smit
The fundamental group of the Hawaiian earring is not free Bart de Smit The fundamental group of the Hawaiian earring is not free (International Journal of Algebra and Computation Vol. 2, No. 1 (1992),
More informationTHE FUNDAMENTAL THEOREM OF ALGEBRA VIA PROPER MAPS
THE FUNDAMENTAL THEOREM OF ALGEBRA VIA PROPER MAPS KEITH CONRAD 1. Introduction The Fundamental Theorem of Algebra says every nonconstant polynomial with complex coefficients can be factored into linear
More informationMEASURE AND INTEGRATION. Dietmar A. Salamon ETH Zürich
MEASURE AND INTEGRATION Dietmar A. Salamon ETH Zürich 12 May 2016 ii Preface This book is based on notes for the lecture course Measure and Integration held at ETH Zürich in the spring semester 2014. Prerequisites
More informationPrime Numbers and Irreducible Polynomials
Prime Numbers and Irreducible Polynomials M. Ram Murty The similarity between prime numbers and irreducible polynomials has been a dominant theme in the development of number theory and algebraic geometry.
More informationMetric Spaces. Chapter 7. 7.1. Metrics
Chapter 7 Metric Spaces A metric space is a set X that has a notion of the distance d(x, y) between every pair of points x, y X. The purpose of this chapter is to introduce metric spaces and give some
More informationRANDOM INTERVAL HOMEOMORPHISMS. MICHA L MISIUREWICZ Indiana University Purdue University Indianapolis
RANDOM INTERVAL HOMEOMORPHISMS MICHA L MISIUREWICZ Indiana University Purdue University Indianapolis This is a joint work with Lluís Alsedà Motivation: A talk by Yulij Ilyashenko. Two interval maps, applied
More informationMICROLOCAL ANALYSIS OF THE BOCHNERMARTINELLI INTEGRAL
PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 00, Number 0, Pages 000 000 S 00029939(XX)00000 MICROLOCAL ANALYSIS OF THE BOCHNERMARTINELLI INTEGRAL NIKOLAI TARKHANOV AND NIKOLAI VASILEVSKI
More informationLet H and J be as in the above lemma. The result of the lemma shows that the integral
Let and be as in the above lemma. The result of the lemma shows that the integral ( f(x, y)dy) dx is well defined; we denote it by f(x, y)dydx. By symmetry, also the integral ( f(x, y)dx) dy is well defined;
More informationThe Ideal Class Group
Chapter 5 The Ideal Class Group We will use Minkowski theory, which belongs to the general area of geometry of numbers, to gain insight into the ideal class group of a number field. We have already mentioned
More informationON SEQUENTIAL CONTINUITY OF COMPOSITION MAPPING. 0. Introduction
ON SEQUENTIAL CONTINUITY OF COMPOSITION MAPPING Abstract. In [1] there was proved a theorem concerning the continuity of the composition mapping, and there was announced a theorem on sequential continuity
More informationFinite dimensional topological vector spaces
Chapter 3 Finite dimensional topological vector spaces 3.1 Finite dimensional Hausdorff t.v.s. Let X be a vector space over the field K of real or complex numbers. We know from linear algebra that the
More information8.1 Examples, definitions, and basic properties
8 De Rham cohomology Last updated: May 21, 211. 8.1 Examples, definitions, and basic properties A kform ω Ω k (M) is closed if dω =. It is exact if there is a (k 1)form σ Ω k 1 (M) such that dσ = ω.
More informationFINITE DIMENSIONAL ORDERED VECTOR SPACES WITH RIESZ INTERPOLATION AND EFFROSSHEN S UNIMODULARITY CONJECTURE AARON TIKUISIS
FINITE DIMENSIONAL ORDERED VECTOR SPACES WITH RIESZ INTERPOLATION AND EFFROSSHEN S UNIMODULARITY CONJECTURE AARON TIKUISIS Abstract. It is shown that, for any field F R, any ordered vector space structure
More informationMathematical Methods of Engineering Analysis
Mathematical Methods of Engineering Analysis Erhan Çinlar Robert J. Vanderbei February 2, 2000 Contents Sets and Functions 1 1 Sets................................... 1 Subsets.............................
More informationTHREE DIMENSIONAL GEOMETRY
Chapter 8 THREE DIMENSIONAL GEOMETRY 8.1 Introduction In this chapter we present a vector algebra approach to three dimensional geometry. The aim is to present standard properties of lines and planes,
More information1 if 1 x 0 1 if 0 x 1
Chapter 3 Continuity In this chapter we begin by defining the fundamental notion of continuity for real valued functions of a single real variable. When trying to decide whether a given function is or
More informationThe BanachTarski Paradox
University of Oslo MAT2 Project The BanachTarski Paradox Author: Fredrik Meyer Supervisor: Nadia S. Larsen Abstract In its weak form, the BanachTarski paradox states that for any ball in R, it is possible
More informationMetric Spaces. Chapter 1
Chapter 1 Metric Spaces Many of the arguments you have seen in several variable calculus are almost identical to the corresponding arguments in one variable calculus, especially arguments concerning convergence
More informationProblem Set. Problem Set #2. Math 5322, Fall December 3, 2001 ANSWERS
Problem Set Problem Set #2 Math 5322, Fall 2001 December 3, 2001 ANSWERS i Problem 1. [Problem 18, page 32] Let A P(X) be an algebra, A σ the collection of countable unions of sets in A, and A σδ the collection
More informationCourse 221: Analysis Academic year , First Semester
Course 221: Analysis Academic year 200708, First Semester David R. Wilkins Copyright c David R. Wilkins 1989 2007 Contents 1 Basic Theorems of Real Analysis 1 1.1 The Least Upper Bound Principle................
More informationSome stability results of parameter identification in a jump diffusion model
Some stability results of parameter identification in a jump diffusion model D. Düvelmeyer Technische Universität Chemnitz, Fakultät für Mathematik, 09107 Chemnitz, Germany Abstract In this paper we discuss
More informationFACTORING POLYNOMIALS IN THE RING OF FORMAL POWER SERIES OVER Z
FACTORING POLYNOMIALS IN THE RING OF FORMAL POWER SERIES OVER Z DANIEL BIRMAJER, JUAN B GIL, AND MICHAEL WEINER Abstract We consider polynomials with integer coefficients and discuss their factorization
More informationTHE DYING FIBONACCI TREE. 1. Introduction. Consider a tree with two types of nodes, say A and B, and the following properties:
THE DYING FIBONACCI TREE BERNHARD GITTENBERGER 1. Introduction Consider a tree with two types of nodes, say A and B, and the following properties: 1. Let the root be of type A.. Each node of type A produces
More informationNotes on metric spaces
Notes on metric spaces 1 Introduction The purpose of these notes is to quickly review some of the basic concepts from Real Analysis, Metric Spaces and some related results that will be used in this course.
More informationIntroduction to Topology
Introduction to Topology Tomoo Matsumura November 30, 2010 Contents 1 Topological spaces 3 1.1 Basis of a Topology......................................... 3 1.2 Comparing Topologies.......................................
More informationAn example of a computable
An example of a computable absolutely normal number Verónica Becher Santiago Figueira Abstract The first example of an absolutely normal number was given by Sierpinski in 96, twenty years before the concept
More informationDouble Sequences and Double Series
Double Sequences and Double Series Eissa D. Habil Islamic University of Gaza P.O. Box 108, Gaza, Palestine Email: habil@iugaza.edu Abstract This research considers two traditional important questions,
More informationA CHARACTERIZATION OF QUASICONVEX VECTORVALUED FUNCTIONS. 1. Introduction
1 2 A CHARACTERIZATION OF QUASICONVEX VECTORVALUED FUNCTIONS Abstract. The aim of this paper is to characterize in terms of scalar quasiconvexity the vectorvalued functions which are Kquasiconvex with
More informationA NOTE ON CLOSED GRAPH THEOREMS. 1. Introduction
Acta Math. Univ. Comenianae Vol. LXXV, 2(2006), pp. 209 218 209 A NOTE ON CLOSED GRAPH THEOREMS F. GACH Abstract. We give a common generalisation of the closed graph theorems of De Wilde and of Popa. 1.
More informationINVARIANT METRICS WITH NONNEGATIVE CURVATURE ON COMPACT LIE GROUPS
INVARIANT METRICS WITH NONNEGATIVE CURVATURE ON COMPACT LIE GROUPS NATHAN BROWN, RACHEL FINCK, MATTHEW SPENCER, KRISTOPHER TAPP, AND ZHONGTAO WU Abstract. We classify the leftinvariant metrics with nonnegative
More information6. Metric spaces. In this section we review the basic facts about metric spaces. d : X X [0, )
6. Metric spaces In this section we review the basic facts about metric spaces. Definitions. A metric on a nonempty set X is a map with the following properties: d : X X [0, ) (i) If x, y X are points
More informationWe call this set an ndimensional parallelogram (with one vertex 0). We also refer to the vectors x 1,..., x n as the edges of P.
Volumes of parallelograms 1 Chapter 8 Volumes of parallelograms In the present short chapter we are going to discuss the elementary geometrical objects which we call parallelograms. These are going to
More informationCHAPTER II THE LIMIT OF A SEQUENCE OF NUMBERS DEFINITION OF THE NUMBER e.
CHAPTER II THE LIMIT OF A SEQUENCE OF NUMBERS DEFINITION OF THE NUMBER e. This chapter contains the beginnings of the most important, and probably the most subtle, notion in mathematical analysis, i.e.,
More informationInvariant Metrics with Nonnegative Curvature on Compact Lie Groups
Canad. Math. Bull. Vol. 50 (1), 2007 pp. 24 34 Invariant Metrics with Nonnegative Curvature on Compact Lie Groups Nathan Brown, Rachel Finck, Matthew Spencer, Kristopher Tapp and Zhongtao Wu Abstract.
More informationInvertible elements in associates and semigroups. 1
Quasigroups and Related Systems 5 (1998), 53 68 Invertible elements in associates and semigroups. 1 Fedir Sokhatsky Abstract Some invertibility criteria of an element in associates, in particular in nary
More informationMathematical Physics, Lecture 9
Mathematical Physics, Lecture 9 Hoshang Heydari Fysikum April 25, 2012 Hoshang Heydari (Fysikum) Mathematical Physics, Lecture 9 April 25, 2012 1 / 42 Table of contents 1 Differentiable manifolds 2 Differential
More informationChapter 15 Introduction to Linear Programming
Chapter 15 Introduction to Linear Programming An Introduction to Optimization Spring, 2014 WeiTa Chu 1 Brief History of Linear Programming The goal of linear programming is to determine the values of
More informationLINEAR ALGEBRA W W L CHEN
LINEAR ALGEBRA W W L CHEN c W W L Chen, 1997, 2008 This chapter is available free to all individuals, on understanding that it is not to be used for financial gain, and may be downloaded and/or photocopied,
More informationt := maxγ ν subject to ν {0,1,2,...} and f(x c +γ ν d) f(x c )+cγ ν f (x c ;d).
1. Line Search Methods Let f : R n R be given and suppose that x c is our current best estimate of a solution to P min x R nf(x). A standard method for improving the estimate x c is to choose a direction
More informationMath212a1010 Lebesgue measure.
Math212a1010 Lebesgue measure. October 19, 2010 Today s lecture will be devoted to Lebesgue measure, a creation of Henri Lebesgue, in his thesis, one of the most famous theses in the history of mathematics.
More informationMINIMAL GENERATOR SETS FOR FINITELY GENERATED SHIFTINVARIANT SUBSPACES OF L 2 (R n )
MINIMAL GENERATOR SETS FOR FINITELY GENERATED SHIFTINVARIANT SUBSPACES OF L 2 (R n ) MARCIN BOWNIK AND NORBERT KAIBLINGER Abstract. Let S be a shiftinvariant subspace of L 2 (R n ) defined by N generators
More information1 Polyhedra and Linear Programming
CS 598CSC: Combinatorial Optimization Lecture date: January 21, 2009 Instructor: Chandra Chekuri Scribe: Sungjin Im 1 Polyhedra and Linear Programming In this lecture, we will cover some basic material
More informationModern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh
Modern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh Peter Richtárik Week 3 Randomized Coordinate Descent With Arbitrary Sampling January 27, 2016 1 / 30 The Problem
More informationImprecise probabilities, bets and functional analytic methods in Łukasiewicz logic.
Imprecise probabilities, bets and functional analytic methods in Łukasiewicz logic. Martina Fedel joint work with K.Keimel,F.Montagna,W.Roth Martina Fedel (UNISI) 1 / 32 Goal The goal of this talk is to
More information4: SINGLEPERIOD MARKET MODELS
4: SINGLEPERIOD MARKET MODELS Ben Goldys and Marek Rutkowski School of Mathematics and Statistics University of Sydney Semester 2, 2015 B. Goldys and M. Rutkowski (USydney) Slides 4: SinglePeriod Market
More informationSimilarity and Diagonalization. Similar Matrices
MATH022 Linear Algebra Brief lecture notes 48 Similarity and Diagonalization Similar Matrices Let A and B be n n matrices. We say that A is similar to B if there is an invertible n n matrix P such that
More informationCHAPTER IV NORMED LINEAR SPACES AND BANACH SPACES
CHAPTER IV NORMED LINEAR SPACES AND BANACH SPACES DEFINITION A Banach space is a real normed linear space that is a complete metric space in the metric defined by its norm. A complex Banach space is a
More informationRi and. i=1. S i N. and. R R i
The subset R of R n is a closed rectangle if there are n nonempty closed intervals {[a 1, b 1 ], [a 2, b 2 ],..., [a n, b n ]} so that R = [a 1, b 1 ] [a 2, b 2 ] [a n, b n ]. The subset R of R n is an
More informationDate: April 12, 2001. Contents
2 Lagrange Multipliers Date: April 12, 2001 Contents 2.1. Introduction to Lagrange Multipliers......... p. 2 2.2. Enhanced Fritz John Optimality Conditions...... p. 12 2.3. Informative Lagrange Multipliers...........
More informationThe Dirichlet Unit Theorem
Chapter 6 The Dirichlet Unit Theorem As usual, we will be working in the ring B of algebraic integers of a number field L. Two factorizations of an element of B are regarded as essentially the same if
More informationHOMEWORK 5 SOLUTIONS. n!f n (1) lim. ln x n! + xn x. 1 = G n 1 (x). (2) k + 1 n. (n 1)!
Math 7 Fall 205 HOMEWORK 5 SOLUTIONS Problem. 2008 B2 Let F 0 x = ln x. For n 0 and x > 0, let F n+ x = 0 F ntdt. Evaluate n!f n lim n ln n. By directly computing F n x for small n s, we obtain the following
More informationLebesgue Measure on R n
8 CHAPTER 2 Lebesgue Measure on R n Our goal is to construct a notion of the volume, or Lebesgue measure, of rather general subsets of R n that reduces to the usual volume of elementary geometrical sets
More informationMath 225A, Differential Topology: Homework 3
Math 225A, Differential Topology: Homework 3 Ian Coley October 17, 2013 Problem 1.4.7. Suppose that y is a regular value of f : X Y, where X is compact and dim X = dim Y. Show that f 1 (y) is a finite
More informationSolutions to Tutorial 2 (Week 3)
THE UIVERSITY OF SYDEY SCHOOL OF MATHEMATICS AD STATISTICS Solutions to Tutorial 2 (Week 3) MATH3969: Measure Theory and Fourier Analysis (Advanced) Semester 2, 2016 Web Page: http://sydney.edu.au/science/maths/u/ug/sm/math3969/
More informationsconvexity, model sets and their relation
sconvexity, model sets and their relation Zuzana Masáková Jiří Patera Edita Pelantová CRM2639 November 1999 Department of Mathematics, Faculty of Nuclear Science and Physical Engineering, Czech Technical
More informationFuzzy Differential Systems and the New Concept of Stability
Nonlinear Dynamics and Systems Theory, 1(2) (2001) 111 119 Fuzzy Differential Systems and the New Concept of Stability V. Lakshmikantham 1 and S. Leela 2 1 Department of Mathematical Sciences, Florida
More informationI. GROUPS: BASIC DEFINITIONS AND EXAMPLES
I GROUPS: BASIC DEFINITIONS AND EXAMPLES Definition 1: An operation on a set G is a function : G G G Definition 2: A group is a set G which is equipped with an operation and a special element e G, called
More informationFIXED POINT SETS OF FIBERPRESERVING MAPS
FIXED POINT SETS OF FIBERPRESERVING MAPS Robert F. Brown Department of Mathematics University of California Los Angeles, CA 90095 email: rfb@math.ucla.edu Christina L. Soderlund Department of Mathematics
More informationMATH 4330/5330, Fourier Analysis Section 11, The Discrete Fourier Transform
MATH 433/533, Fourier Analysis Section 11, The Discrete Fourier Transform Now, instead of considering functions defined on a continuous domain, like the interval [, 1) or the whole real line R, we wish
More informationORIENTATIONS. Contents
ORIENTATIONS Contents 1. Generators for H n R n, R n p 1 1. Generators for H n R n, R n p We ended last time by constructing explicit generators for H n D n, S n 1 by using an explicit nsimplex which
More informationALMOST COMMON PRIORS 1. INTRODUCTION
ALMOST COMMON PRIORS ZIV HELLMAN ABSTRACT. What happens when priors are not common? We introduce a measure for how far a type space is from having a common prior, which we term prior distance. If a type
More informationThe Heat Equation. Lectures INF2320 p. 1/88
The Heat Equation Lectures INF232 p. 1/88 Lectures INF232 p. 2/88 The Heat Equation We study the heat equation: u t = u xx for x (,1), t >, (1) u(,t) = u(1,t) = for t >, (2) u(x,) = f(x) for x (,1), (3)
More informationGeometry of Linear Programming
Chapter 2 Geometry of Linear Programming The intent of this chapter is to provide a geometric interpretation of linear programming problems. To conceive fundamental concepts and validity of different algorithms
More informationx if x 0, x if x < 0.
Chapter 3 Sequences In this chapter, we discuss sequences. We say what it means for a sequence to converge, and define the limit of a convergent sequence. We begin with some preliminary results about the
More informationStochastic Inventory Control
Chapter 3 Stochastic Inventory Control 1 In this chapter, we consider in much greater details certain dynamic inventory control problems of the type already encountered in section 1.3. In addition to the
More informationMATH10212 Linear Algebra. Systems of Linear Equations. Definition. An ndimensional vector is a row or a column of n numbers (or letters): a 1.
MATH10212 Linear Algebra Textbook: D. Poole, Linear Algebra: A Modern Introduction. Thompson, 2006. ISBN 0534405967. Systems of Linear Equations Definition. An ndimensional vector is a row or a column
More informationContinuity of the Perron Root
Linear and Multilinear Algebra http://dx.doi.org/10.1080/03081087.2014.934233 ArXiv: 1407.7564 (http://arxiv.org/abs/1407.7564) Continuity of the Perron Root Carl D. Meyer Department of Mathematics, North
More informationSection 1.1. Introduction to R n
The Calculus of Functions of Several Variables Section. Introduction to R n Calculus is the study of functional relationships and how related quantities change with each other. In your first exposure to
More informationCHAPTER III  MARKOV CHAINS
CHAPTER III  MARKOV CHAINS JOSEPH G. CONLON 1. General Theory of Markov Chains We have already discussed the standard random walk on the integers Z. A Markov Chain can be viewed as a generalization of
More information