arXiv subjects
Publications and source records attributed to Saugata Basu.
We prove an analog of the Szemerédi-Trotter theorem in the plane for definable curves and points in any o-minimal structure over an arbitrary real closed field $\mathrm{R}$. One new ingredient in the proof is an extension of the well known crossing number inequality for graphs to the case of embeddings in any o-minimal structure over an arbitrary real closed field.
Let $\mathrm{R}$ be a real closed field and $\mathrm{D} \subset \mathrm{R}$ an ordered domain. We consider the algorithmic problem of computing the generalized Euler-Poincaré characteristic of real algebraic as well as semi-algebraic subsets of $\mathrm{R}^k$, which are defined by symmetric polynomials with coefficients in $\mathrm{D}$. We give algorithms for computing the generalized Euler-Poincaré characteristic of such sets, whose complexities measured by the number the number of arithmetic operations in $\mathrm{D}$, are polynomially bounded in terms of $k$ and the number of polynomials in the input, assuming that the degrees of the input polynomials are bounded by a constant. This is in contrast to the best complexity of the known algorithms for the same problems in the non-symmetric situation, which are singly exponential. This singly exponential complexity for the latter problem is unlikely to be improved because of hardness result ($\#\mathbf{P}$-hardness) coming from discrete complexity theory.
We consider symmetric (under the action of products of finite symmetric groups) real algebraic varieties and semi-algebraic sets, as well as symmetric complex varieties in affine and projective spaces, defined by polynomials of degrees bounded by a fixed constant $d$. We prove that if a Specht module, $\mathbb{S}^λ$, appears with positive multiplicity in the isotypic decomposition of the cohomology modules of such sets, then the rank of the partition $λ$ is bounded by $O(d)$. This implies a polynomial (in the dimension of the ambient space) bound on the number of such modules. Furthermore, we prove a polynomial bound on the multiplicities of those that do appear with positive multiplicity in the isotypic decomposition of the above mentioned cohomology modules. We give some applications of our methods in proving lower bounds on the degrees of defining polynomials of certain symmetric semi-algebraic sets, as well as improved bounds on the Betti numbers of the images under projections of (not necessarily symmetric) bounded real algebraic sets, improving in certain situations prior results of Gabrielov, Vorobjov and Zell.
We derive a formula expressing the average number $E_n$ of real lines on a random hypersurface of degree $2n-3$ in $\mathbb{R}\textrm{P}^n$ in terms of the expected modulus of the determinant of a special random matrix. In the case $n=3$ we prove that the average number of real lines on a random cubic surface in $\mathbb{R}\textrm{P}^3$ equals: $$E_3=6\sqrt{2}-3.$$ Our technique can also be used to express the number $C_n$ of complex lines on a generic hypersurface of degree $2n-3$ in $\mathbb{C}\textrm{P}^n$ in terms of the determinant of a random Hermitian matrix. As a special case we obtain a new proof of the classical statement $C_3=27.$ We determine, at the logarithmic scale, the asymptotic of the quantity $E_n$, by relating it to $C_n$ (whose asymptotic has been recently computed D. Zagier). Specifically we prove that: $$\lim_{n\to \infty}\frac{\log E_n}{\log C_n}=\frac{1}{2}.$$ Finally we show that this approach can be used to compute the number $R_n=(2n-3)!!$ of real lines, counted with their intrinsic signs, on a generic real hypersurface of degree $2n-3$ in $\mathbb{R}\textrm{P}^n$.
Let $\mathrm{R}$ be a real closed field, and $\mathrm{D} \subset \mathrm{R}$ an ordered domain. We describe an algorithm that given as input a polynomial $P \in \mathrm{D} [ X_{1},\ldots,X_{k} ]$, and a finite set, $\mathcal{A}= \{ p_{1}, \ldots,p_{m} \}$, of points contained in $V= \mathrm{Zer}( P, \mathrm{R}^{k})$ described by real univariate representations, computes a roadmap of $V$ containing $\mathcal{A}$. The complexity of the algorithm, measured by the number of arithmetic operations in $\mathrm{D} $ is bounded by $\left( \sum_{i=1}^{m} D^{O ( \log^{2} ( k ) )}_{i} +1 \right) ( k^{\log ( k )} d )^{O ( k\log^{2} ( k ))}$, where $d= \mathrm{deg} ( P )$, and $D_{i}$ is the degree of the real univariate representation describing the point $p_{i}$. The best previous algorithm for this problem had complexity $\mathrm{card} ( \mathcal{A} )^{O ( 1 )} d^{O ( k^{3/2} )}$ due to Basu, Roy, Safey-El-Din, and Schost (2012), where it is assumed that the degrees of the polynomials appearing in the representations of the points in $\mathcal{A}$ are bounded by $d^{O ( k )}$. As an application of our result we prove that for any real algebraic subset $V$ of $\mathbb{R}^{k}$ defined by a polynomial of degree $d$, any connected component $C$ of $V$ contained in the unit ball, and any two points of $C$, there exist a semi-algebraic path connecting them in $C$, of length at most $( k ^{\log (k )} d )^{O ( k\log ( k ) )}$, consisting of at most $( k ^{\log (k )} d )^{O ( k\log ( k ) )}$ curve segments of degrees bounded by $( k ^{\log ( k )} d )^{O ( k \log ( k) )}$. While it was known previously, by a result of D'Acunto and Kurdyka, that there always exists a path of length $( O ( d ) )^{k-1}$ connecting two such points, there was no upper bound on the complexity of such a path.
Let $\mathrm{R}$ be a real closed field. The problem of obtaining tight bounds on the Betti numbers of semi-algebraic subsets of $\mathrm{R}^k$ in terms of the number and degrees of the defining polynomials has been an important problem in real algebraic geometry with the first results due to Ole{\uı}nik and Petrovski{\uı}, Thom and Milnor. These bounds are all exponential in the number of variables $k$. Motivated by several applications in real algebraic geometry, as well as in theoretical computer science, where such bounds have found applications, we consider in this paper the problem of bounding the equivariant Betti numbers of symmetric algebraic and semi-algebraic subsets of $\mathrm{R}^k$. We obtain several asymptotically tight upper bounds. In particular, we prove that if $S\subset \mathrm{R}^k$ is a semi-algebraic subset defined by a finite set of $s$ symmetric polynomials of degree at most $d$, then the sum of the $\mathfrak{S}_k$-equivariant Betti numbers of $S$ with coefficients in $\mathbb{Q}$ is bounded by $(skd)^{O(d)}$. Unlike the classical bounds on the ordinary Betti numbers of real algebraic varieties and semi-algebraic sets, the above bound is polynomial in $k$ when the degrees of the defining polynomials are bounded by a constant. As an application we improve the best known bound on the ordinary Betti numbers of the projection of a compact algebraic set improving for any fixed degree the best previously known bound for this problem due to Gabrielov, Vorobjov and Zell.
We prove graded bounds on the individual Betti numbers of affine and projective complex varieties. In particular, we give for each $p,d,r$, explicit bounds on the $p$-th Betti numbers of affine and projective subvarieties of $\mathrm{C}^k$, $\mathbb{P}^k_{\mathrm{C}}$, as well as products of projective spaces, defined by $r$ polynomials of degrees at most $d$ as a function of $p,d$ and $r$. Unlike previous bounds these bounds are independent of $k$, the dimension of the ambient space. We also prove as consequences of our technique certain homological and representational stability results for sequences of complex projective varieties which could be of independent interest. Finally, we highlight differences in computational complexities of the problem of computing Betti numbers of complex as opposed to real projective varieties.
Let $\mathrm{R}$ be a real closed field and $Q_1, \ldots, Q_{\ell} \in \mathrm{R}[X_1, \ldots,X_k]$ such that for each $i, 1 \leq i \leq \ell$, $\mathrm{deg} (Q_i) \leq d_i$. For $1 \leq i \leq \ell$, denote by $\mathcal{Q}_i = \{Q_1, \ldots, Q_i \}$, $V_i$ the real variety defined by $\mathcal{Q}_i$, and $k_i$ an upper bound on the real dimension of $V_i$ (by convention $V_0 = \mathrm{R}^k$ and $k_0 = k$). Suppose also that \[ 2 \leq d_1 \leq d_2 \leq \frac{1}{k + 1} d_3 \leq \frac{1}{(k + 1)^2} d_4 \leq \cdots \leq \frac{1}{(k + 1)^{\ell - 3}} d_{\ell - 1} \leq \frac{1}{(k + 1)^{\ell - 2}} d_{\ell}, \] and that $\ell \leq k$. We prove that the number of semi-algebraically connected components of $V_{\ell}$ is bounded by \[ O (k)^{2 k} \left(\prod_{1 \leq j < \ell} d_j^{k_{j - 1} - k_j} \right) d_{\ell}^{k_{\ell - 1}}. \] This bound can be seen as a weak extension of the classical Bezout inequality (which holds only over algebraically closed fields and is false over real closed fields) to varieties defined over real closed fields. Additionally, if $\mathcal{P} \subset \mathrm{R}[X_1, \ldots, X_k]$ is a finite family of polynomials with $\mathrm{deg} (P) \leq d$ for all $P \in \mathcal{P}$, $\mathrm{card}( \mathcal{P}) = s$, and $d_{\ell} \leq \frac{1}{k + 1} d$, we prove that the number of semi-algebraically connected components of the realizations of all realizable sign conditions of the family $\mathcal{P}$ restricted to $V_{\ell}$ is bounded by \[ O (k)^{2 k} (s d)^{k_{\ell}} \left(\prod_{1 \leq j \leq \ell} d_j^{k_{j - 1} - k_j} \right). \]
We present a polynomial partitioning theorem for finite sets of points in the real locus of an irreducible complex algebraic variety of codimension at most two. This result generalizes the polynomial partitioning theorem on the Euclidean space of Guth and Katz, and its extension to hypersurfaces by Zahl and by Kaplan, Matoušek, Sharir and Safernová. We also present a bound for the number of incidences between points and hypersurfaces in the four-dimensional Euclidean space. It is an application of our partitioning theorem together with the refined bounds for the number of connected components of a semi-algebraic set by Barone and Basu.
Let $K \subset {\mathbb R}^n$ be a compact definable set in an o-minimal structure over $\mathbb R$, e.g., a semi-algebraic or a subanalytic set. A definable family $\{ S_δ|\> 0< δ\in {\mathbb R} \}$ of compact subsets of $K$, is called a monotone family if $S_δ\subset S_η$ for all sufficiently small $δ> η>0$. The main result of the paper is that when $\dim K \le 2$ there exists a definable triangulation of $K$ such that for each (open) simplex $Λ$ of the triangulation and each small enough $δ>0$, the intersection $S_δ\cap Λ$ is equivalent to one of the five standard families in the standard simplex (the equivalence relation and a standard family will be formally defined). The set of standard families is in a natural bijective correspondence with the set of all five lex-monotone Boolean functions in two variables. As a consequence, we prove the two-dimensional case of the topological conjecture in [6] on approximation of definable sets by compact families. We introduce most technical tools and prove statements for compact sets $K$ of arbitrary dimensions, with the view towards extending the main result and proving the topological conjecture in the general case.
We survey both old and new developments in the theory of algorithms in real algebraic geometry -- starting from effective quantifier elimination in the first order theory of reals due to Tarski and Seidenberg, to more recent algorithms for computing topological invariants of semi-algebraic sets. We emphasize throughout the complexity aspects of these algorithms and also discuss the computational hardness of the underlying problems. We also describe some recent results linking the computational hardness of decision problems in the first order theory of the reals, with that of computing certain topological invariants of semi-algebraic sets. Even though we mostly concentrate on exact algorithms, we also discuss some numerical approaches involving semi-definite programming that have gained popularity in recent times.
Let $\mathrm{R}$ be a real closed field and $\mathrm{D} \subset \mathrm{R}$ an ordered domain. We give an algorithm that takes as input a polynomial $Q \in \mathrm{D}[X_1,\ldots,X_k]$, and computes a description of a roadmap of the set of zeros, $\mathrm{Zer}(Q,\mathrm{R}^k)$, of $Q$ in $\mathrm{R}^k$. The complexity of the algorithm, measured by the number of arithmetic operations in the ordered domain $\mathrm{D}$, is bounded by $d^{O(k \sqrt{k})}$, where $d = \mathrm{deg}(Q)\ge 2$. As a consequence, there exist algorithms for computing the number of semi-algebraically connected components of a real algebraic set, $\mathrm{Zer}(Q,\mathrm{R}^k)$, whose complexity is also bounded by $d^{O(k \sqrt{k})}$, where $d = \mathrm{deg}(Q)\ge 2$. The best previously known algorithm for constructing a roadmap of a real algebraic subset of $\mathrm{R}^k$ defined by a polynomial of degree $d$ has complexity $d^{O(k^2)}$.
We consider sets and maps defined over an o-minimal structure over the reals, such as real semi-algebraic or subanalytic sets. A {\em monotone map} is a multi-dimensional generalization of a usual univariate monotone function, while the closure of the graph of a monotone map is a generalization of a compact convex set. In a particular case of an identically constant function, such a graph is called a {\em semi-monotone set}. Graphs of monotone maps are, generally, non-convex, and their intersections, unlike intersections of convex sets, can be topologically complicated. In particular, such an intersection is not necessarily the graph of a monotone map. Nevertheless, we prove a Helly-type theorem, which says that for a finite family of subsets of $\Real^n$, if all intersections of subfamilies, with cardinalities at most $n+1$, are non-empty and graphs of monotone maps, then the intersection of the whole family is non-empty and the graph of a monotone map.
In [S. Basu, A. Gabrielov, N. Vorobjov, Semi-monotone sets. arXiv:1004.5047v2 (2011)] we defined semi-monotone sets, as open bounded sets, definable in an o-minimal structure over the reals, and having connected intersections with all translated coordinate cones in R^n. In this paper we develop this theory further by defining monotone functions and maps, and studying their fundamental geometric properties. We prove several equivalent conditions for a bounded continuous definable function or map to be monotone. We show that the class of graphs of monotone maps is closed under intersections with affine coordinate subspaces and projections to coordinate subspaces. We prove that the graph of a monotone map is a topologically regular cell. These results generalize and expand the corresponding results obtained in Basu et al. for semi-monotone sets.
We prove that the number of distinct homotopy types of limits of one-parameter semi-algebraic families of closed and bounded semi-algebraic sets is bounded singly exponentially in the additive complexity of any quantifier-free first order formula defining the family. As an important consequence, we derive that the number of distinct homotopy types of semi-algebraic subsets of $\mathbb{R}^k$ defined by a quantifier-free first order formula $Φ$, where the sum of the additive complexities of the polynomials appearing in $Φ$ is at most $a$, is bounded by $2^{(k+a)^{O(1)}}$. This proves a conjecture made by Basu and Vorobjov [On the number of homotopy types of fibres of a definable map, J. Lond. Math. Soc. (2) 2007, 757--776].
We prove that toric cubes, which are images of $[0,1]^d$ under monomial maps, are the closures of graphs of monotone maps, and in particular semi-algebraically homeomorphic to closed balls.
Let $\R$ be a real closed field, $\mathcal{P},\mathcal{Q} \subset \R[X_1,...,X_k]$ finite subsets of polynomials, with the degrees of the polynomials in $\mathcal{P}$ (resp. $\mathcal{Q}$) bounded by $d$ (resp. $d_0$). Let $V \subset \R^k$ be the real algebraic variety defined by the polynomials in $\mathcal{Q}$ and suppose that the real dimension of $V$ is bounded by $k'$. We prove that the number of semi-algebraically connected components of the realizations of all realizable sign conditions of the family $\mathcal{P}$ on $V$ is bounded by $$ \displaylines{\sum_{j=0}^{k'}4^j{s +1\choose j}F_{d,d_0,k,k'}(j),}$$ where $s = \card \; \mathcal{P}$, and $$F_{d,d_0,k,k'}(j)= \textstyle\binom{k+1}{k-k'+j+1} \;(2d_0)^{k-k'}d^j\; \max{2d_0,d}^{k'-j} +2(k-j+1) .$$ In case $2 d_0 \leq d$, the above bound can be written simply as $$ \displaylines{\sum_{j = 0}^{k'} {s+1 \choose j}d^{k'} d_0^{k-k'} O(1)^{k} = (sd)^{k'} d_0^{k-k'} O(1)^k} $$ (in this form the bound was suggested by J. Matousek. Our result improves in certain cases (when $d_0 \ll d$) the best known bound of $$ \sum_{1 \leq j \leq k'} \binom{s}{j} 4^{j} d(2d-1)^{k-1} $$ on the same number proved earlier in the case $d=d_0$. The distinction between the bound $d_0$ on the degrees of the polynomials defining the variety $V$ and the bound $d$ on the degrees of the polynomials in $\mathcal{P}$ that appears in the new bound is motivated by several applications in discrete geometry.
Toda \cite{Toda} proved in 1989 that the (discrete) polynomial time hierarchy, $\mathbf{PH}$, is contained in the class $\mathbf{P}^{#\mathbf{P}}$, namely the class of languages that can be decided by a Turing machine in polynomial time given access to an oracle with the power to compute a function in the counting complexity class $#\mathbf{P}$. This result, which illustrates the power of counting is considered to be a seminal result in computational complexity theory. An analogous result (with a compactness hypothesis) in the complexity theory over the reals (in the sense of Blum-Shub-Smale real machines \cite{BSS89}) was proved in \cite{BZ09}. Unlike Toda's proof in the discrete case, which relied on sophisticated combinatorial arguments, the proof in \cite{BZ09} is topological in nature in which the properties of the topological join is used in a fundamental way. However, the constructions used in \cite{BZ09} were semi-algebraic -- they used real inequalities in an essential way and as such do not extend to the complex case. In this paper, we extend the techniques developed in \cite{BZ09} to the complex projective case. A key role is played by the complex join of quasi-projective complex varieties. As a consequence we obtain a complex analogue of Toda's theorem. The results contained in this paper, taken together with those contained in \cite{BZ09}, illustrate the central role of the Poincaré polynomial in algorithmic algebraic geometry, as well as, in computational complexity theory over the complex and real numbers -- namely, the ability to compute it efficiently enables one to decide in polynomial time all languages in the (compact) polynomial hierarchy over the appropriate field.