Darnes Vilariño Ayala

Tiancalli06: An Agent for the Supply Chain Management Game 2006

For two years now, the Tiancalli project has been developed. This project consists in the construction of a TAC SCM compliant agent. The contest is about different software agents competing for a simulated market. This year, we present the Tiancalli06 agent. In this article the learning techniques implemented to the agent are described. These strategies are used so the agent can compete in the market and obtain incomes. The final results about the performance of the agent on the three phases that it participated are also presented - Qualification, Seeding and Quarter Finals. The previous results of this Project can be found in [8, 9, 10, 11, 12].

Statistic Analysis for the Tiancalli Agents on TAC SCM 2005 and 2006

During the last two years we have developed the Tiancalli project. This project includes the construction of TAC SCM compliant agents. The TAC SCM is a contest about different software agents competing in a simulated market that has been organized for four years by Trading Agent Research Group. In order to find the strength and weakness of our agent, we have made a statistic research with both 2005 and 2006 game results. The objective of this article is to present the results obtained in this comparison. Also it is important to discuss a more concrete basis about factors that must be modified, such as the ones which should be kept on the TAC SCM 2007

Agents Development for a Supply Chain Management Game

With the purpose of studying dynamic markets where different suppliers and customers interact, a platform which allows competition between agents in an environment of supply chains has been developed at Carnegie Mellon University. It is known as The trading agent competition: supply chain management game (TAC SCM). In this game, agents that represent computer sellers use different strategies to gain customer requests and to assure that the required components to satisfy them would be received in time. At the end of the game, the winner agent gets the highest utilities through the non-static conditions of the market. In this article, eight different agents able to compete in this game are presented and the results of their performance in different rounds are discussed

BUAP: A Recursive Approach to the Data-Centric Track of INEX 2011

A recursive approach for keyword search on XML data for the Ad-Hoc Search Task of INEX 2011 is presented in this paper. The aim of this approach was to detect the concrete part (in the representation tree) of the XML document containing the expected answer. For this purpose, we initially obtain a tree structure, which represents an XML document, tagged by levels. A typical search engine based on posting lists is used in order to determine those documents that match in some degree with the terms appearing in the given query(topic). Thereafter, in a recursively process, we navigate into the tree structure until we find the best match for the topic. The obtained results are shown and compared with the best overall submission score obtained in the competition.

The Virtual Paragraph as a Retrieval Information Technique Implanted in Mobile Agents

Nowadays the information contained in networks increases constantly, and the way to access this information and to gather important data to the users becomes a tedious task. The purpose of this work is to show the use of mobile agents to recover valuable information for the users they represent. In this work mobile agents are endowed with different strategies to retry information. Our approach is to obtain summaries by using the transition point technique (Punto de Transition) and to select those documents and sentences that have a high similarity with the users requirements. To validate the summaries, the DUC 2002 and a statistical analysis are used. Moreover, the implementation of mobile agents is made in JADE.

Mobile Agents with ProActive for Document Extracts

Information retrieval (IR) deals with the searching of document extracts that provide the nearest approach to a users query. Due the growing of Internet there is a necessity to effectively use the information available in remote servers. Here data are distributed and, therefore, the searching mechanisms should be more efficient; otherwise, this may increase the traffic in a network. One of the benefits of mobile agents is that they reduce the communication costs between the hosts where the information is available and the central processor where the processing of information is done. That is, instead of sending messages with all the data in a local host, mobile agents are able to work in local processors and to send only the data that they consider relevant. This work shows how mobile agents can facilitate the process of information retrieval by using a platform to program distributed systems (ProActive). Different migration strategies for agents are implemented and the results are showed

Comparing four Methods to Select Keywords that Use n-Grams to Generate Summaries

In this paper an algorithm to generate document extracts is proposed. This extract is composed of the most significative sentences of the current document. These sentences are chosen by considering their similitude with a so-denominated virtual paragraph (VP). The VP is formed by the key words that are obtained by the four methods that use n-grams, employed in this research. The experiments are performed over 100 documents of the DUC-2000 collection.