Marcin Paprzycki

Developing a Model Agent-based E-Commerce System

It is easy to realize that goals set behind a large class of agent systems match these put forward for systems defined as e-service intelligence. In this chapter we describe a model agent-based e-commerce system that utilizes rule-based approach for price negotiations. Furthermore, the proposed system attempts at mediating the apparent contradiction between agent mobility and intelligence.

Real stock trading using soft computing models

The main focus of this study is to compare different performances of soft computing paradigms for predicting the direction of individuals stocks. Three different artificial intelligence techniques were used to predict the direction of both Microsoft and Intel stock prices over a period of thirteen years. We explore the performance of artificial neural networks trained using backpropagation and conjugate gradient algorithm and a Mamdani and Takagi Sugeno fuzzy inference system learned using neural learning and genetic algorithm. Once all the different models were built the last part of the experiment was to determine how much profit can be made using these methods versus a simple buy and hold technique.

tilizing agent teams in grid resource management preliminary considerations

Recently it was suggested that (mobile) software agents can provide an infrastructure for resource management in grids. In this note we introduce an approach based on agent teams, and discuss how it can he used in grid resource management. Details of initial implementation of one of its functionalities are discussed

A system for ensuring data integrity in grid environments

Data integrity has to become one of the central concerns of large-scale distributed computing systems such as the grid, whose primary products are the results of computation. In order to maintain the integrity of this data, the system must be resilient to diverse attacks and tampering. The system should also encourage positive influences on its integrity in addition to discouraging or eliminating negative ones. In this paper we develop a model of trust for grid participants based on the use of reputation systems and associated feedback mechanisms.

Data mining approach for analyzing call center performance

One of the difficulties of using Artificial Neural Networks (ANNs) to estimate atmospheric temperature is the large number of potential input variables available. In this study, four different feature extraction methods were used to reduce the input vector to train four networks to estimate temperature at different atmospheric levels. The four techniques used were: genetic algorithms (GA), coefficient of determination (CoD), mutual information (MI) and simple neural analysis (SNA). The results demonstrate that of the four methods used for this data set, mutual information and simple neural analysis can generate networks that have a smaller input parameter set, while still maintaining a high degree of accuracy.

Extending Maple to the grid: design and implementation

One of the important issues facing the development of the grid as the computational framework of the future is availability of grid-enabled software. In this context, we discuss possible approaches to constructing a grid-enabled version of a computer algebra system. Our case study involves Maple: the proposed Maple2g package allows the connection between Maple and the computational grids based on the Globus Toolkit. We present the design of the Maple2g package and follow with a thorough discussion of its implementation.

Knowledge Representation in the Agent-Based Travel Support System

One of the interesting problems in implementing an e-commerce system is the question of what knowledge is required and how it should be implemented. In this note we attempt to address this problem in the guise of an agent-based e-commerce system for providing personalized travel services.

Design and implementation of a grid extension for Maple

One of the important issues facing the development of the grid as a computational framework of the future is availability of grid-enabled software. In this context, we discuss possible approaches to constructing a grid-enabled version of a computer algebra system. Our case study involves Maple: the proposed Maple2g package allows the connection between Maple and the computational grids based on the Globus Toolkit. We present the design of the Maple2g package and follow with a detailed discussion of its implementation. Finally, we illustrate performance of Maple2g in a number of experiments.

Towards a Grid-Aware Computer Algebra System

One of the developments that can lead to a wider usage of grid technologies is grid-enabling of application software, among them computer algebra systems. A case study described here involves Maple. The proposed maple2g package allows the connection between the current version of Maple and the computational grid based on Globus.

A grid-based parallel maple

Popularity and success of computational grids will depend, among others, on the availability of application software. Maple2g is a grid-oriented extension of Maple. One of its components allows the access of grid services within Maple, while another one use of multiple computational units. The latter component is discussed in this paper. It is based on a master-slave paradigm, and it is implemented using Globus Toolkit GT3, mpiJava and MPICH-G2. Preliminary experiments are reported and discussed. These are proving that a reasonable time reduction of computational-intensive applications written in Maple can be obtained by using multiple kernels running on different grid sites.