A birth–death model for Web cache systems: Numerical solutions and simulation

In this paper, we study Web cache hit rates by introducing a birth–death model. A system consisting of a single Web server with a single cache stores Web pages that are classified as hot pages (popular pages) and cold pages (less popular pages). Given requested probabilities for each class, the stochastic model provides the mean hit rate for a random replacement algorithm and the upper and low bounds for other algorithms. Numerical results from the analysis are validated using the output of simulation programs that utilize the LRU algorithm.     

A fixed-point algorithm to estimate the Yule-Simon distribution parameter

The Yule-Simon distribution is a discrete probability distribution related to preferential attachment processes such as the growth in the number of species per genus in some higher taxon of biotic organisms, the distribution of the sizes of cities, the wealth distribution among individuals, the number of links to pages in the World Wide Web, among others. In this paper we present an algorithm to, given a set of observations stemmed from a Yule process, obtain the parameter of the Yule-Simon distribution with maximum likelihood. In order to test our algorithm, we use a modified Polya urn process simulation to generate some data that was used as input to our algorithm. We make a comparison of our algorithm with other methods and also we show an application to some empirical data.     

Multiple factor hierarchical clustering algorithm for?large scale web page and search engine clickstream data

The developments in World Wide Web and the advances in digital data collection and storage technologies during the last two decades allow companies and organizations to store and share huge amounts of electronic documents. It is hard and inefficient to manually organize, analyze and present these documents. Search engine helps users to find relevant information by present a list of web pages in response to queries. How to assist users to find the most relevant web pages from vast text collections efficiently is a big challenge. The purpose of this study is to propose a hierarchical clustering method that combines multiple factors to identify clusters of web pages that can satisfy users?? information needs. The clusters are primarily envisioned to be used for search and navigation and potentially for some form of visualization as well. An experiment on Clickstream data from a processional search engine was conducted to examine the results shown that the clustering method is effective and efficient, in terms of both objective and subjective measures.     

A study of the similarities between topics

This paper presents a new methodology whereby the relationships among points of interest appearing in Internet Web pages may be graphically represented. This is done by quantifying the various points and representing them graphically using an intuitive interpretation of the similarity of sets. The similarity or interrelationships existing among various lines of research may be studied based on information taken from the Web pages using function taken from Learning Theory. A visual analysis of those interrelationships has been applied to the field of machine learning during the time interval May 2003–May 2004.     

Optimal Web site reorganization considering information overload and search depth

Adaptive Web sites have been extensively studied as the Internet and Web sites continuously grow. Based on Web usage mining, most of these studies focus on providing assistance to users instead of optimizing the Web site itself. This study proposes 0–1 programming models for reorganizing Web sites based on the cohesion between Web pages obtained by Web usage mining. Additionally, the proposed models reduce the information overload and search depth for users surfing in the Web site. Additionally, a heuristic approach was proposed to reduce the required computation time. The proposed models were verified and extensively tested with numerical examples.     

How to Pack Trees

In a virtual memory system, the address space is partitioned into pages, and the main memory serves as a cache to the disk. In this setting, we address the following problem: Given a tree, find an allocation of its nodes to pages, so-called a packing, which optimizes the cache performance for some access pattern to the tree nodes. We investigate a model for tree access in which a node is accessed only via the path leading to it from the root. Two cost functions are considered: the total number of different pages visited in the search, and the number of page faults incurred. It is shown that both functions can be optimized simultaneously. An efficient dynamic programming algorithm to find an optimal packing is presented. The problem of finding an optimal packing which also uses the minimum number of pages is shown to be NP-complete. However, an efficient approximation algorithm is presented. This algorithm finds a packing that uses the minimum number of pages and requires at most one extra page fault per search. Finally, we study this problem in the context of dynamic trees which allow insertions and deletions.     

Predicting Web Users' Next Access Based on Log Data

This article considers models that describe how people browse the Web. We restrict our attention to navigation patterns within a single site, and base our study on standard Web server access logs. Given a visitor's previous activities on the site, we propose models that predict their next page request. If the prediction is reasonably accurate, we might consider “prefetching” the page before the visitor requests it. A more conservative use for such predictions would be to simply update the freshness records in a proxy or network cache, eliminating unnecessary If-Modified-Since requests. Using data from the Web site for the Computing and Mathematical Sciences Research Division of Lucent Technologies (cm.bell-labs.com) we first evaluate the predictive performance of low-order Markov models. We next consider mixtures of first-order Markov models, achieving a kind of clustering of Web pages in the site. This approach is shown to perform well, while significantly reducing the space required to store the model. Finally, we explore a Bayesian approach using a Dirichlet prior on the collection of links available to a user at each stage in their travels through the site. We show that the posterior probabilities derived under this model are fairly close to the cross-validation estimates of the probability of success.     

Google PageRank as mean playing time for pinball on the reverse web

It is known that the output from Google’s PageRank algorithm may be interpreted as (a) the limiting value of a linear recurrence relation that is motivated by interpreting links as votes of confidence, and (b) the invariant measure of a teleporting random walk that follows links except for occasional uniform jumps. Here, we show that, for a sufficiently frequent jump rate, the PageRank score may also be interpreted as a mean finishing time for a reverse random walk. At a general step this new process either (i) remains at the current page, (ii) moves to a page that points to the current page, or (iii) terminates. The process is analogous to a game of pinball where a ball bounces between pages before eventually dropping down the exit chute. This new interpretation of PageRank gives another view of the principle that highly ranked pages will be those that are linked into by highly ranked pages that have relatively few outgoing links.     

基于多层次映射和遗传策略的语义Web搜索方法

随着语义Web概念的提出及相关技术的发展,基于语义的Web信息检索逐渐成为语义Web研究的热点.首先分析了传统的Web搜索面临的诸多挑战,然后设计了一种新的基于遗传语义搜索算法和和多层次映射的Web搜索方法.该方法充分结合遗传算法并行搜索方式,能扩大搜索空间,快速实现全局优化.实验结果表明该算法能有效提高Web搜索的质量和效率.     

基于中心平面的聚类模型及在电商中的应用

因为k-平面聚类算法(kPC)和k-中心平面聚类算法(kPPC)构建的聚类中心平面是无限延伸的,这会影响聚类的性能,所以提出了局部的k-中心平面聚类(L-kPPC)算法.此算法在kPPC中引入了k-均值聚类算法(k-mean),这样使得样本点都聚集在类中心周围.L-kPPC利用了各聚类中心平面的局部特征构建类中心平面,...     

Convex interval interpolation using a three-term staircase algorithm

Motivated by earlier considerations of interval interpolation problems as well as a particular application to the reconstruction of railway bridges, we deal with the problem of univariate convexity preserving interval interpolation. To allow convex interpolation, the given data intervals have to be in (strictly) convex position. This property is checked by applying an abstract three-term staircase algorithm, which is presented in this paper. Additionally, the algorithm provides strictly convex ordinates belonging to the data intervals. Therefore, the known methods in convex Lagrange interpolation can be used to obtain interval interpolants. In particular, we refer to methods based on polynomial splines defined on grids with additional knots. Received September 22, 1997 / Revised version received May 26, 1998     

Visualizing geometric algorithms over the Web

The visual nature of geometry applications makes them a natural area where visualization can be an effective tool for demonstrating algorithms. In this paper we propose a new model, called Mocha, for interactive visualization of algorithms over the World Wide Web. Mocha is a distributed model with a client-server architecture that optimally partitions the software components of a typical algorithm execution and visualization system, and leverages the power of the Java language, which has become the standard for distributing interactive platform-independent applications across the Web. Mocha provides high levels of security, protects the algorithm code, places a light communication load on the Internet, and allows users with limited computing resources to access executions of computationally expensive algorithms. The user interface combines fast responsiveness with the powerful authoring capabilities of hypertext narratives.We describe the architecture of Mocha, show its advantages over previous methods, and present a prototype that can be accessed by any user with a Java-enabled Web browser. The Mocha prototype has been widely accessed over the Web, as demonstrated by the statistics that we have collected, and the Mocha model has been adopted by other research groups. Mocha is currently part of a broader system, called GeomNet, which performs distributed geometric computing over the Internet.     

The book crossing number of a graph

Let G be a graph on n vertices and m edges. The book crossing number of G is defined as the minimum number of edge crossings when the vertices of G are placed on the spine of a k-page book and edges are drawn on pages, such that each edge is contained by one page. Our main results are two polynomial time algorithms to generate near optimal drawing of G on books. The first algorithm give an O(log² n) times optimal solution, on small number of pages, under some restrictions. This algorithm also gives rise to the first polynomial time algorithm for approximating the rectilinear crossing number so that the coordinates of vertices in the plane are small integers, thus resolving a recent open question concerning the rectilinear crossing number. Moreover, using this algorithm we improve the best known upper bounds on the rectilinear crossing number. The second algorithm generates a drawing of G with O(m²/k²) crossings on k pages. This is within a constant multiplicative factor from our general lower bound of Ω(m³/n²k²), provided that m = Ψ(n²). © 1996 John Wiley & Sons, Inc.     

A Combined Global & Local Search (CGLS) Approach to Global Optimization

This paper presents a general approach that combines global search strategies with local search and attempts to find a global minimum of a real valued function of n variables. It assumes that derivative information is unreliable; consequently, it deals with derivative free algorithms, but derivative information can be easily incorporated. This paper presents a nonmonotone derivative free algorithm and shows numerically that it may converge to a better minimum starting from a local nonglobal minimum. This property is then incorporated into a random population to globalize the algorithm. Convergence to a zero order stationary point is established for nonsmooth convex functions, and convergence to a first order stationary point is established for strictly differentiable functions. Preliminary numerical results are encouraging. A Java implementation that can be run directly from the Web allows the interested reader to get a better insight of the performance of the algorithm on several standard functions. The general framework proposed here, allows the user to incorporate variants of well known global search strategies. Research done under the cooperation agreement between Universidade de Vigo and Universidad Simón Bolívar.     

Flexible job shop scheduling with overlapping in operations

In this paper, flexible job shop scheduling problem with a new approach, overlapping in operations, is discussed. In many flexible job shops, a customer demand can be released more than one for each job, where demand determines the quantity of each finished job ordered by a customer. In these models each job has a demand more than one. This assumption is an important and practical issue for many flexible job shops such as petrochemical industries. To consider this assumption, we use a new approach, named overlapping in operations. In this approach, embedded operations of each job can be performed due to overlap considerations in which each operation may be overlapped with the others because of its nature. The overlapping is limited by structural constraints, such as the dimensions of the box to be packed or the capacity of the container used to move the pieces from one machine to the next. Since this problem is well known as NP-Hard class, a hierarchical approach used simulated annealing algorithm is developed to solve large problem instances. Moreover, a mixed integer linear programming (MILP) method is presented. To evaluate the validity of the proposed SA algorithm, the results are compared with the optimal solution obtained with the traditional optimization technique (The Branch and Bound method). The computational results validate the efficiency and effectiveness of the proposed algorithm. Also the computational results show that the overlapping considering can improve the makespan and machines utilization measures. So the proposed algorithm can be applied easily in real factory conditions and for the large size problems and it should thus be useful to both practitioners and researchers.     

An infeasible (exterior point) simplex algorithm for assignment problems

The so called Modified Hung—Rom Algorithm, based upon theoretical considerations of Hirsch-paths, seems to be one of the most efficient algorithms for assignment problems. Since any two basic feasible solutions to a linear problem can always be connected with a short simplex path passing through the infeasible region, development of algorithms based upon theoretical considerations on infeasible paths seems to be of great practical interest. This paper presents an algorithm of this kind for assignment problems.     

树状网络上的Web代理服务器最优放置问题

一般网络上Web代理服务器(Web proxy)最优放置问题是一个NP困难问题.此文讨论树状网络上的最优放置问题,改进了已有结果,得到了一个时间复杂度为O(nhk)的多项式时间算法,这里n为网络结点数,h为树的高度,而k为要放置的代理服务器个数.     

An adaptive improvement on PageRank algorithm

In this article, we introduce the Google’s method for quality ranking of web page in a formal mathematical format, use the power iteration to improve the PageRank, and also discuss the effect of different q to the PageRank, as well as how a PageRank will be changed if more links are added to one page or removed from some pages.