Pilu Crescenzi

On-line algorithms for the channel assignment problem in cellular networks

We consider the on-line channel assignment problem in the case of cellular networks and we formalize this problem as an on-line load balancing problem for temporary tasks with restricted assignment. For the latter problem, we provide a general solution (denoted as the cluster algorithm ) and we characterize its competitive ratio in terms of the combinatorial properties of the graph representing the network. We then compare the cluster algorithm with the greedy one when applied to the channel assignment problem: it turns out that the competitive ratio of the cluster algorithm is strictly better than the competitive ratio of the greedy algorithm. The cluster method is general enough to be applied to other on-line load balancing problems and, for some topologies, it can be proved to be optimal.

On computing the diameter of real-world undirected graphs

We propose a new algorithm for the classical problem of computing the diameter of undirected unweighted graphs, namely, the maximum distance among all the pairs of nodes, where the distance of a pair of nodes is the number of edges contained in the shortest path connecting these two nodes. Although its worst-case complexity is O(nm) time, where n is the number of nodes and m is the number of edges of the graph, we experimentally show that our algorithm works in O(m) time in practice, requiring few breadth-first searches to complete its task on almost 200 real-world graphs.

Fully integrating algorithm visualization into a cs2 course.: a two-year experience

We describe a two-year experience of fully integrating algorithm visualization technology into a CS2 course on data structures and algorithms. Our integration methodology was based on the engagement taxonomy proposed by the working group on Improving the Educational Impact of Algorithm Visualization: in particular, we used five forms of engagement of this taxonomy, that is, the no-viewing, the viewing, the changing, the constructing and the presenting forms. The integration of algorithm visualization technology into the course culminated in the writing of a textbook on the design, analysis and visualization of data structures and algorithms, whose reading is strictly dependent on the use of an algorithm visualization tool, called Alvie, which has been developed by the authors.

MOMOSE: a mobility model simulation environment for mobile wireless ad-hoc networks

This paper describes MOMOSE, a highly flexible and easily extensible environment for the simulation of mobility models. MOMOSE not only allows a programmer to easily integrate a new mobility model into the set of models already included in its distribution, but it also allows the user to let the nodes of the MANET move in different ways by associating any mobility model to any subset of the nodes themselves. Moreover, MOMOSE can be easily adapted in order to record, during the simulation time, all the data necessary for the evaluation of the performance of any communication protocol or of any MANET-based application.

Spatial node distribution of manhattan path based random waypoint mobility models with applications

In this paper, we study the spatial node stationary distribution of two variations of the Random Waypoint (in short, RWP) mobility model. In particular, differently from the RWP mobility model, that connects source to destination points by straight lines, our models make use of Manhattan or (more realistically) Bezier paths. We provide analytical results for the spatial node stationary distribution for the two Manhattan based RWP mobility models and experimental evidence that the Bezier based models do not significantly differ from the Manhattan ones. This implies that Manhattan based RWP models can be considered a good approximation of the more realistic Bezier ones. As a case study, we exploit our results about one of the two Manhattan based RWP models to derive an upper bound on the transmission range of the nodes of a MANET, moving according to this model, that with high probability guarantees the connectivity of the communication graph.

Integrating test generation functionality into the Teaching Machine environment

We propose an extension of the Teaching Machine project, called Quiz Generator, that allows instructors to produce assessment quizzes in the field of algorithm and data structures quite easily. This extension makes use of visualization techniques and is based on new features of the Teaching Machine that allow third-party visualizers to be added as plugins and for new scripting capabilities. Using these new capabilities, several quiz types have already been produced, which can be applied to any algorithm and/or data structure for which the necessary visualizer plugins exist.

Performance Evaluation of a Chord-Based JXTA Implementation

In this paper, we propose a new implementation of the JXTA framework, based on the integration of the original JXTA version with the Chord DHT management protocol. The resulting new JXTA implementation is experimentally compared with the original one, by referring to the JXTA performance model proposed by Halepovic et al. in 2004: the results give strong evidence to the fact that the JXTA rendezvous protocol performances can be improved up to one order of magnitude by replacing the original loosely-consistent DHT with the real Chord DHT.

Optimal covering designs: complexity results and new bounds

In this paper we investigate the problem of computing optimal lottery schemes. From a computational complexity point of view, we prove that the variation of this problem in which the sets to be covered are specified in the input is log |I|-approximable (where I denotes the collection of sets to be covered) and it cannot be approximated within a factor smaller than log |I|, unless P = NP. From a combinatorial point of view, we propose new constructions based on the combination of the partitioning technique and of known results regarding the construction of sets of coverings. By means of this combination we will be able to improve several upper bounds on the cardinality of optimal lottery schemes.

A tool to develop electronic course books based on WWW technologies, resources and usability criteria

An electronic course book (ECB in short) is a learning module consisting of hyperdocuments with a functional use of interactivity and multimedia, presented on the WWW and/or CDROM [1]. In this paper we propose an ECB producer application, which can assist any author in the development of an ECB based on WWW usability criteria and which presents a collection of several multimedia elements which can enhance the process of learning and which differentiate an electronic course book from a classical paper book. The ECB, which represents an electronic form of classroom support, will turn out to be useful both to teachers, since they will take advantage of the slide-based presentation of the text and of the several simulation tools included in the ECB, and to students, since they will be able to learn by reading, by doing and by answering.

On-line load balancing made simple: Greedy strikes back

We provide a new approach to the on-line load balancing problem in the case of restricted assignment of temporary weighted tasks. The approach is very general and allows us to derive on-line algorithms whose competitive ratio is characterized by some combinatorial properties of the underlying graph G representing the problem: in particular, the approach consists in applying the greedy algorithm to a suitably constructed subgraph of G. In the paper, we prove the NP-hardness of the problem of computing an optimal or even a c-approximate subgraph, for some constant c>1. Nevertheless, we show that, for several interesting problems, we can easily compute a subgraph yielding an optimal on-line algorithm. As an example, the effectiveness of this approach is shown by the hierarchical server model introduced by Bar-Noy et al. (2001). In this case, our method yields simpler algorithms whose competitive ratio is at least as good as the existing ones. Moreover, the algorithm analysis turns out to be simpler. Finally, we give a sufficient condition for obtaining, in the general case, O(n)-competitive algorithms with our technique: this condition holds in the case of several problems for which a @W(n) lower bound is known.