Stochastic decomposition for retrial queues

Summary This paper deals with the stochastic decomposition property for retrial queues. This property is connected with similar results for vacation models. As applications, the moments of the number of customers in orbit and the rate of convergence under high retrial intensity can be obtained. This work was supported under Grant PR161/93-4777     

Some logical aspects of information processing

This paper is concerned with dynamic aspects of knowledge representation. A model for information processing is introduced which is based on acyclic directed graphs with starting points. This approach is combined with ideas from non-monotonic reasoning.     

An Inventory Model with an Inventory-Level-Dependent Demand Rate and Relaxed Terminal Conditions

This paper investigates an inventory system with an inventory-level-dependent demand rate followed by a constant demand rate. The analysis previously conducted on this model imposed a terminal condition of zero inventory at the end of the order cycle. This condition is not warranted for systems with inventory-level-dependent demand rates, even for deterministic models. This paper relaxes this restriction, and an appropriate model is developed.     

A variational approach to stochastic nonlinear parabolic problems

This work is concerned with an optimal control approach to stochastic nonlinear parabolic diffusion equations with monotonically increasing nonlinearity. This approach leads to sharper existence and uniqueness results under minimal growth conditions on nonlinear diffusion coefficients.     

Fractional behaviour of partial differential equations whose coefficients are exponential functions of the space variable

There exists a close link between fractional systems and infinite dimensional systems described by diffusion equations. This link can be demonstrated analytically and is reminded in this article. This fractional behaviour results in fact in the system infinite dimension along with constant geometric characteristics. This article demonstrates that several other classes of differential equations also exhibit, on a frequency band, a fractional behaviour. The fractional behaviour is obtained with these equations on a space of finite dimension but with particular geometric characteristics.     

An improved class of estimators of a finite population quantile in sample surveys

This work proposes a general class of estimators for a finite population quantile using auxiliary information. This information is provided by the population means of auxiliary variables. The optimum estimator in this class is derived. This result is supported with a numerical example.     

The multifit algorithm for set partitioning containing kernels

This pager investigates the set partitioning containing kernels. This problem can alsobe considered as the identical machine scheduling prohlem with nonsimultaneous machine release times. That the algorithm MULTIFIT has a worst ease bound of 6/5 is proved. Throughcombining MULTIFIT and LPT. an algorithm MULTILPT with a worat case bound of 7/6 has been obtained.     

A combinatorial generalization of polytopes

Primoids and duoids are collections of subsets of a fixed finite set with a natural generalization of a pivoting property of convex polytopes. This structure is precisely what is necessary for the application of complementary pivoting algorithms. This paper investigates the combinatorial structure of primoids and duoids, showing them to form the circuits and cocircuits of a binary matroid. This matroid is then compared with the simplicial geometries of Crapo and Rota.     

Chaos in the one-dimensional wave equation

This paper deals with the chaotic behavior of the solutions of a mixed problem for the one-dimensional wave equation with a quadratic boundary condition. This behavior is studied through the connection between the energy function and quadratic discrete dynamical systems.     

Hill climbing on a sample function of a Gaussian Markov process

This note deals with a theoretical procedure for hill climbing on a sample function of a Gaussian Markov process.This work was performed while the author was with the Raytheon Company, Research Division, Waltham, Massachusetts.     

Asymptotic behavior of gradient systems with small time delays

This paper is concerned with the asymptotic behavior of gradient systems with small time delays. Roughly speaking, we show that each bounded solution will approach one of the equilibria in case the equilibria are isolated. This indicates that gradient systems with small time delays behave very much like the nondelayed ones.     

Weighted switching lifted digital redesign with state saturation for cascaded analogue systems

This paper presents the newly developed digital redesign controlscheme with a weighted switching strategy and lifted digitalcontrol, for a cascaded analogue system with state saturation.This method improves the generally poor transition responsecaused by state-saturation constraints. It replaces an existingor designed analogue controller with state saturation, by adigital controller with almost identical performance.     

Directional derivative of the value function in parametric optimization

This paper presents with elementary proofs some results on the directional derivative of the optimal value of a finite dimensional optimization problem with parameters.This research was supported by the Natural Sciences and Engineering Research Council of Canada under Grant A-9273.     

Independence numbers of graphs and generators of ideals

This article investigates the generators of certain homogeneous ideals which are associated with graphs with bounded independence numbers. These ideals first appeared in the theory oft-designs. The main theorem suggests a new approach to the Clique Problem which isNP-complete. This theorem has a more general form in commutative algebra dealing with ideals associated with unions of linear varieties. This general theorem is stated in the article; a corollary to it generalizes Turán’s theorem on the maximum graphs with a prescribed clique number. Research supported in part by NSF Grant MCS77-03533.     

带重点的广义Cauchy矩阵及其相关问题

本文详细讨论了带重点的广义Cauchy矩阵的位移结构、快速求逆公式、可逆性判别条件及其与矩阵有理切插值问题的关系,并给出了求解此类矩阵的线性方程组的快速算法,推广了文[1]和[2]的结果.     

Truncated Distributive Lattices: Conceptual Structures of Simple-Implicational Theories

Logical relationships in everyday human thought are predominantly inferences with one-element premises. This becomes apparent in the practice of Formal Concept Analysis by the frequent occurrence of truncated distributive lattices as concept lattices. This paper gives a mathematization of the underlying everyday theories of logical relationships and elaborates useful mathematical results, in particular about algorithmically drawing concept lattices which correspond to the everyday logical theories. This revised version was published online in June 2006 with corrections to the Cover Date.     

一种Tandem网络模型

本文考虑的是一个MAP输入的tandem队列，运用一个新颖的方法把它模拟成一个带爆炸块依赖阶数的拟生灭过程，这样使得我们可以很好地分析它的联合时间分布和平稳状态的停留时间分布布。     

Improved tabu search algorithm for the handling of route duration constraints in vehicle routing problems with time windows

This note introduces a refinement to a previously proposed tabu search algorithm for vehicle routing problems with time windows. This refinement yields new best known solutions on a set of benchmark instances of the multi-depot, the periodic and the site-dependent vehicle routing problems with time windows.     

Learning recursive probability trees from probabilistic potentials

A Recursive Probability Tree (RPT) is a data structure for representing the potentials involved in Probabilistic Graphical Models (PGMs). This structure is developed with the aim of capturing some types of independencies that cannot be represented with previous structures. This capability leads to improvements in memory space and computation time during inference. This paper describes a learning algorithm for building RPTs from probability distributions. The experimental analysis shows the proper behavior of the algorithm: it produces RPTs encoding good approximations of the original probability distributions.