Andrzej Kasprzak

The web replica allocation and topology assignment problem in wide area networks: algorithms and computational results

This paper studies the problem of designing wide area networks. In the paper the web replica allocation and topology assignment problem with budget constraint is considered. The goal is select web replica allocation at nodes, network topology, channel capacities and flow routes in order to minimize average delay per packet and web replica connecting cost at nodes subject to budget constraint. The problem is NP-complete. Then, the branch and bound method is used to construct the exact algorithm. Also an approximate algorithm is presented. Some computational results are reported. Based on computational experiments, several properties of the considered problem are formulated.

Dynamic routing of anycast and unicast traffic in Elastic Optical Networks

In this paper, we focus on issues related to dynamic routing of anycast and unicast traffic demands in Elastic Optical Networks (EONs). The growing interest in cloud computing and content-oriented services triggers the need to examine anycast traffic in the context of the EON concept, which is perceived as a promising approach for future optical transport networks. Anycasting - defined as one-to-one of many transmission - ideally corresponds to concepts of cloud computing and content-oriented services, where the same service/content is available in many, geographically spread data centers. In the paper, we propose two types of algorithms for dynamic routing of anycast and unicast requests. The former algorithm is based on the shortest path first (SPF) method, while the latter algorithm is based on the selection of a routing path from a set of candidate paths (CP). To examine performance of the algorithms, we run simulations using a real pan-European topology and traffic patterns created according to Cisco predictions. The results show a tradeoff between SPF and CP methods in terms of the blocking probability and execution time. Moreover, we observe that the blocking probability depends on the number of data centers placed in the network.

Isozyme pattern of pyruvate kinase during hepatocarcino-genesis induced by 2-acetylaminofluorene in rat liver☆

Abstract The biosynthesis of pyruvate kinase isozymes has been studied in rat liver during the administration of carcinogen 2 -acetylaminofluorene. Transition from a predominance of isozyme L to K was observed during hepatocarcinogenesis. Changes in isozymic pattern and specific activity of pyruvate kinase were observed before histological alterations typical for cancer formation in rat liver could be detected. The dedifferentiation during hepatocarcinogenesis results in switching on the gene for isozyme K and switching off the gene for isozyme L. Within a single liver cell, however, both genes are not functioning simultaneously. Metabolic consequences of the observed changes are discussed.

Active Learning Classification of Drifted Streaming Data

Abstract Objects being recognized may arrive continuously to a classifier in the form of data stream, therefore contemporary classification systems have to make a decision not only on the basis of the static data, but on the data in motion as well. Additionally, we would like to start a classifier exploitation as soon as possible, then the models which can improve their models during exportation are very desirable. Basically, we may produce the model on the basis a few learning objects only and then we use and improve the classifier when new data comes. This concept is still vibrant and may be used in the plethora of practical cases. Nevertheless, constructing such a system we should realize, that we have the limited resources (as memory and computational power) at our disposal. Additionally, during the exploitation of a classifier system the chosen characteristic of the classifier model may change within a time. This phenomena is called concept drift and may lead the deep deterioration of the classification performance. This work deals with the data stream classification with the presence of concept drift . We propose a novel classifier training algorithm based on the sliding windows approach, which allows us to implement forgetting mechanism, i.e., that old objects come from outdated model will not be taken into consideration during the classifier updating and on the other hand we assume that only part of arriving examples can be labeled, because we assume that we have a limited budget for labeling. We will employ active learning paradigm to choose an “interesting” objects to be be labeled. The proposed approach has been evaluated on the basis of the computer experiments carried out on the data streams. Obtained results confirmed the usability of proposed method to the smoothly drifted data stream classification.

Exact and approximate algorithms for topological design of wide area networks with non-simultaneous single commodity flows

The problem is concerned with designing of wide area computer networks. This problem is very important from practical point of view because the network, when designed efficiently, can and does become the lifeblood of a business organization. The problem consists in selection of the nonsimultaneous single commodity flow routes, channel capacities and the wide area network topology in order to maximize the value of the min-max flow subject to the budget constraint. The problem is NP-complete. Then, the branch and bound method is used to construct the exact algorithm. Several properties of the problem are formulated and proved. Computational results are reported.

An Exact Algorithm for Host Allocation, Capacity and Flow Assignment Problem in WAN

The paper presents the exact algorithm for capacity and flow assignment and hosts allocation problem in wide area computer networks. The problem consists in selection of flow routes, channel capacities and hosts allocation at nodes in order to minimize the total average delay per packet and the connecting cost of the hosts at nodes subject to budget constraint. The branch and bound method to construct the exact algorithm for solving the considered problem is used.

Evaluation of strip nesting algorithms: An experimentation system for the practical users

Strip nesting problems are well known to be of NP-hard complexity, even in simple cases. A multitude of algorithms now exist that aim at solving such problems efficiently. The experimentation system described here helps users with industrial applications to decide which algorithm suits their requirements best. As an illustration, it is applied to three algorithms, based on: simulated annealing (SA), tabu search (TS), and ant colony optimization (ACO). The system can be treated as an input-output system, in which the parameters of the problem and the algorithms for solving the problem are the inputs, and the introduced measures of algorithms’ quality are the outputs. Experiments are conducted in accordance with a two-stage approach. At the first stage, the tuning process for each algorithm is made. At the second stage, the efficiencies of the optimized algorithms are compared. In the paper, the results of the comparison obtained with the benchmark databases are discussed, and conclusions concerning efficiency of the considered algorithms are drawn accordingly.

Weighted Aging Classifier Ensemble for the Incremental Drifted Data Streams

Evolving systems are recently focus of intense research because for most of the real problems we can observe that the parameters of the decision tasks should adapt to new conditions. In classification such a problem is usually called concept drift. The paper deals with the data stream classification where we assume that the concept drift is sudden but its rapidity is limited. To deal with this problem we propose a new algorithm called Weighted Aging Ensemble WAE, which is able to adapt to changes of classification model parameters. The method is inspired by well-known algorithm Accuracy Weighted Ensemble AWE which allows to change the line-up of a classifier ensemble, but the proposed method incudes two important modifications: i classifier weights depend on the individual classifier accuracies and time they have been spending in the ensemble, ii individual classifier are chosen to the ensemble on the basis on the non-pairwise diversity measure. The proposed method was evaluated on the basis of computer experiments which were carried out on SEA dataset. The obtained results encourage us to continue the work on the proposed concept.

An experimentation system for testing bee behavior based algorithm to solving a transportation problem

This paper describes a system, called eTransport to solve a transportation problem. The system was applied in experiments that test meta-heuristic algorithms for solving such optimization task. The engine of this system is based on a bee behavior based algorithm, called KAPI, designed by the authors. In order to make a comparison of the results obtained by KAPI, the Ants Colony System and Tabu Search algorithms were also applied. The eTransport simulator can generate real-life scenarios on Google Maps, while configuring and running tested algorithms, and finally, displaying the solutions found. Some illustrative examples of experiments are presented and discussed. The analysis of results of multi-scenario simulations shows the advantages of the KAPI algorithm, and justifies the conclusion that KAPI is much better and more effective at finding solutions to problems of this kind than other known algorithms.

Evaluation and Comparison of Task Allocation Algorithms for Mesh Networks

The paper concerns the task allocation problem for dynamic mesh structured systems. Three allocation algorithms have been evaluated, including well-known Full Search (FS), and Window-Stack Based Algorithm (WSBA) as well as Random Node Check (RNC) designed by the authors. The objective of the paper was a comparison of algorithms efficiencies on the basis of the results of simulations made using multi platform experimentation system designed and implemented in Python environment. The reported investigations have shown that the WSBA algorithm seems to be the most promising.