Search

We show that there are planar graphs that require four pages in any book embedding., To be published in Journal of Combinatorial Theory, Series B

We resolve the complexity of revenue-optimal deterministic auctions in the unit-demand single-buyer Bayesian setting, i.e., the optimal item pricing problem, when the buyer's values for the items are independent. We show that the problem of computing a revenue-optimal pricing can be solved in polynomial time for distributions of support size 2, and its decision version is NP-complete for distributions of support size 3. We also show that the problem remains NP-complete for the case of identical distributions.

We begin by observing that (discrete-time) Quasi-Birth-Death Processes (QBDs) are equivalent, in a precise sense, to probabilistic 1-Counter Automata (p1CAs), and both Tree-Like QBDs (TL-QBDs) and Tree-Structured QBDs (TS-QBDs) are equivalent to both probabilistic Pushdown Systems (pPDSs) and Recursive Markov Chains (RMCs). We then proceed to exploit these connections to obtain a number of new algorithmic upper and lower bounds for central computational problems about these models. Our main result is this: for an arbitrary QBD, we can approximate its termination probabilities (i.e., its $G$ matrix) to within $i$ bits of precision (i.e., within additive error $1/2^i$), in time polynomial in \underline{both} the encoding size of the QBD and in $i$, in the unit-cost rational arithmetic RAM model of computation. Specifically, we show that a decomposed Newton's method can be used to achieve this. We emphasize that this bound is very different from the well-known ``linear/quadratic convergence'' of numerical analysis, known for QBDs and TL-QBDs, which typically gives no constructive bound in terms of the encoding size of the system being solved. In fact, we observe (based on recent results) that for the more general TL-QBDs such a polynomial upper bound on Newton's method fails badly. Our upper bound proof for QBDs combines several ingredients: a detailed analysis of the structure of 1-counter automata, an iterative application of a classic condition number bound for errors in linear systems, and a very recent constructive bound on the performance of Newton's method for strongly connected monotone systems of polynomial equations. We show that the quantitative termination decision problem for QBDs (namely, ``is $G_{u,v} \geq 1/2$?'') is at least as hard as long standing open problems in the complexity of exact numerical computation, specifically the square-root sum problem. On the other hand, it follows from our earlier results for RMCs that any non-trivial approximation of termination probabilities for TL-QBDs is sqrt-root-sum-hard.

Trade-off (aka Pareto) curves are typically used to represent the trade-off among different objectives in multiobjective optimization problems. Although trade-off curves are exponentially large for typical combinatorial optimization problems (and infinite for continuous problems), it was observed in Papadimitriou and Yannakakis [On the approximability of trade-offs and optimal access of web sources, in: Proc. 41st IEEE Symp. on Foundations of Computer Science, 2000] that there exist polynomial size e approximations for any e > 0, and that under certain general conditions, such approximate e-Pareto curves can be constructed in polynomial time. In this paper we seek general-purpose algorithms for the efficient approximation of trade-off curves using as few points as possible. In the case of two objectives, we present a general algorithm that efficiently computes an e-Pareto curve that uses at most 3 times the number of points of the smallest such curve; we show that no algorithm can be better than 3-competitive in this setting. If we relax e to any e' > e, then we can efficiently construct an e'-curve that uses no more points than the smallest e-curve. With three objectives we show that no algorithm can be c-competitive for any constant c unless it is allowed to use a larger e value. We present an algorithm that is 4-competitive for any e' > (1 + e)2 - 1. We explore the problem in high dimensions and give hardness proofs showing that (unless P=NP) no constant approximation factor can be achieved efficiently even if we relax e by an arbitrary constant.

Scenario-based specifications such as message sequence charts (MSC) offer an intuitive and visual way to describe design requirements. MSC-graphs allow convenient expression of multiple scenarios, and can be viewed as an early model of the system that can be subjected to a variety of analyses. Problems such as LTL model checking are undecidable for MSC-graphs in general, but are known to be decidable for the class of bounded MSC-graphs.Our first set of results concerns checking realizability of bounded MSC-graphs. An MSC-graph is realizable if there is a distributed implementation that generates precisely the behaviors in the graph. There are two notions of realizability, weak and safe, depending on whether or not we require the implementation to be deadlock-free. It is known that for a finite set of MSCs, weak realizability is coNP-complete while safe realizability has a polynomial-time solution. We establish that for bounded MSC-graphs, weak realizability is, surprisingly, undecidable, while safe realizability is in EXPSPACE.Our second set of results concerns verification of MSC-graphs. While checking properties of a graph G, besides verifying all the scenarios in the set L(G) of MSCs specified by G, it is desirable to verify all the scenarios in the set Lw(G)—the closure of G, that contains the implied scenarios that any distributed implementation of G must include. For checking whether a given MSC M is a possible behavior, checking M∈L(G) is NP-complete, but checking M∈Lw(G) has a quadratic solution. For temporal logic specifications, considering the closure makes the verification problem harder: while checking LTL properties of L(G) is PSPACE-complete for bounded graphs G, checking even simple “local” properties of Lw(G) is undecidable.

A (2 - 2/k) approximation algorithm is presented for the node multiway cut problem, thus matching the result of Dahlhaus et al. (SIAM J. Comput. 23 (4) (1994) 864-894) for the edge version of this problem. This is done by showing that the associated LP-relaxation always has a half-integral optimal solution.

We revisit the issue of the complexity of database queries, in the light of the recent parametric refinement of complexity theory. We show that, if the query size (or the number of variables in the query) is considered as a parameter, then the relational calculus and its fragments (conjunctive queries, positive queries) are classified at appropriate levels of the so-called W hierarchy of Downey and Fellows. These results strongly suggest that the query size is inherently in the exponent of the data complexity of any query evaluation algorithm, with the implication becoming stronger as the expressibility of the query language increases. On the positive side, we show that this exponential dependence can be avoided for the extension of acyclic queries with ≠ (but not

We characterize precisely the complexity of several natural computational problems in NP, which have been proposed but not categorized satisfactorily in the literature: Computing the Vapnik–Chervonenkis dimension of a 0–1 matrix; finding the minimum dominating set of a tournament; satisfying a Boolean expression by perturbing the default truth assignment; and several others. These problems can be solved innO(logn)time, and thus, they are probably not NP-complete. We define two new complexity classes between P and NP, very much in the spirit of MAXNP and MAXSNP. We show that computing the V–C dimension is complete for the more general class, while the other two problems are complete for the weaker class.

We present an algorithm for verifying that a model M with timing constraints satisfies a given temporal property T. The model M is given as a parallel composition of ω-automata Pi, where each automaton Pi is constrained by bounds on delays. The property T is given as an ω-automaton as well, and the verification problem is posed as a language inclusion question L(M) ⊆ L(T). In constructing the composition M of the constrained automata Pi, one needs to rule out the behaviors that are inconsistent with the delay bounds, and this step is (provably) computationally expensive. We propose an iterative solution which involves generating successive approximations Mj to M, with containment L(M) ⊆ L(Mj) and monotone convergence L(Mj) → L(M) within a bounded number of steps. As the succession progresses, the approximations Mj become more complex. At any step of the iteration one may get a proof or a counter-example to the original language inclusion question. The described algorithm is implemented into the verifier Cospan. We illustrate the benefits of our strategy through some examples.

We present a 3/4 polynomial time approximation algorithm for the Maximum Satisfiability problem: Given a set of clauses, find a truth assignment that satisfies the maximum number of clauses. The algorithm applies to the weighted case as well, and involves nontrival application of network flow techniques.

We define a natural variant of NP, MAX NP, and also a subclass called MAX SNP. These are classes of optimization problems, and in fact contain several natural, well-studied ones. We show that problems in these classes can be approximated with some bounded error. Furthermore, we show that a number of common optimization problems are complete for MAX SNP under a kind of careful transformation (called L-reduction) that preserves approximability. It follows that such a complete problem has a polynomial-time approximation scheme iff the whole class does. These results may help explain the lack of progress on the approximability of a host of optimization problems.

Using a pair of finite-state automata to model the transmitter-receiver protocol in a data communications system, we derive lower bounds on the size of automata needed to achieve reliable communication across an error-prone channel. We also show that, at the cost of increasing the size of the automata, a transmission rate close to the theoretical maximum can be achieved.

We discuss the problems of finding maximum and connected maximum k-colorable subgraphs in chordal graphs. We prove that the problems are polynomially solvable when k is fixed and NP-hard when k is not fixed. As a special case, we can find in polynomial time the maximum induced tree and forest of a chordal graph.

We consider the family of graph problems called node-deletion problems, defined as follows; For a fixed graph property Π, what is the minimum number of nodes which must be deleted from a given graph so that the resulting subgraph satisfies Π? We show that if Π is nontrivial and hereditary on induced subgraphs, then the node-deletion problem for Π is NP-complete for both undirected and directed graphs.

We present an algorithm that generates all maximal independent sets of a graph in lexicographic order, with only polynomial delay between the output of two successive independent sets. We also show that there is no polynomial-delay algorithm for generating all maximal independent sets in reverse lexicographic order, unless P=NP.

The following issues in computational complexity remain imprecisely understood: the striking difference in the complexities of computing the permanent and determinant of a matrix despite their similar looking formulae, the complexity of checking if a directed graph contains an even length cycle, and the complexity of computing the number of perfect matchings in a graph using Pfaffian orientations. Via polynomial time equivalences, we show interrelationships among these issues.

A database schema is independent with respect to a given set of constraints if the constraints can be enforced separately in the relations. A polynomial time algorithm is presented that recognizes independent schemas, when the given constraints consist of functional dependencies and the join dependency of the database schema.