Patricia Azevedo Tedesco

Recent advances on multi-agent patrolling

Patrolling is a complex multi-agent task, which usually requires agents to coordinate their decision-making in order to achieve optimal performance of the group as a whole. In previous work, many patrolling strategies were developed, based on different approaches: heuristic agents, negotiation mechanisms, reinforcement learning techniques, techniques based on graph-theory and others. In this paper, we complement these studies by comparing all the approaches developed so far for this domain, using patrolling problem instances that were not dealt with before (i.e. new map topologies). The final results constitute a valuable benchmark for this domain, as well as a survey of the strategies developed so far for this task.

Designing context-sensitive systems: An integrated approach

Context-sensitive systems (CSS) are computer systems that use context to provide more relevant services or information to support users performing their tasks, where context is any information that can be used to characterize the situation in which something exists or occurs. CSS demand that designers consider new aspects and challenges in comparison to traditional applications. In a preliminary experiment, we observed that developers find it difficult to include the concept of context in their applications. However, only few approaches offer integrated domain-independent support on developing CSS. This paper presents an integrated approach to assist the design of CSS. The originality of this work lies on the proposed way of thinking about context, on the proposed context metamodel and on the specification of a process for designing CSS. The metamodel supports building context models by making explicit the concepts related to context manipulation and by separating the context structure model from the CSS behavior model. The design process details the main activities related to context specification and the design of CSS, providing a systematic way to execute these tasks. This work also advances the state of the art related to the understanding of the concept of context (and its associated concepts). Three experimental studies were conducted to evaluate the proposal: its instantiation in the design of a context-sensitive Expert Recommender System, its usage by distinct designers on their CSS projects, and a qualitative evaluation of the overall proposal by experienced CSS developers. These studies showed a good acceptance of our approach and indicated the feasibility of using it on real projects.

Towards an ontology for context representation in groupware

An important issue in groupware is how to improve interaction and collaboration among participants. Through the analysis of the context a user is in or the context that surrounds an interaction, groupware systems can provide users with useful information in that situation. A relevant issue when using context is how to represent context information. Ontologies constitute an interesting method for representing context, since they enable information sharing and reuse. They can also be used by existing inference machines to reason about various contexts. In this paper we propose an ontology to formally represent context in groupware systems. We also present an example where this ontology is used by a logic-based reasoning mechanism for tool recommendation based on the current context of group members. We believe that this ontology could help to understand the role of context in collaboration and thus make the development of context-aware groupware systems easier.

Investigating the specifics of contextual elements management: the CEManTIKA approach

In times where users need to process an ever increasing amount of information to perform more complex tasks in less time, the introduction of context in computer systems is becoming a necessity. However, building a context-sensitive system entails more work in comparison to traditional systems development: in the former, one must care for context-related tasks, such as the acquisition, processing, storage, manipulation and presentation of contextual elements. Context management proposes the separation of context related tasks from the applications business features. This paper presents a study on context management and discusses our proposal for a context manager, called CEManTIKA (Contextual Elements Management Through Incremental Knowledge Acquisition). A scenario is presented to illustrate its applicability.

Negotiator agents for the patrolling task

Multi-agent systems can be used to perform patrolling tasks in various domains. In this work, we compare the results obtained by new negotiation based approaches with previous ones. By splitting the nodes of the world graph, the negotiator agents reduce the path they have to walk and the number of nodes to patrol, making it easier to maintain a low average idleness in world nodes. Auctions are the negotiation mechanisms used to split the nodes of the world, the agents bid on nodes based in their utility function. Empirical evaluation has shown the effectiveness of this distributed approach, as the results obtained are substantially better than those previously achieved by negotiator agents. The agent types presented in this work are more scalable and reactive since they can perform patrolling in worlds of all sizes and topology types. Besides, they are more stable as indicated by the low standard deviation obtained in node idleness.

The Gravitational Strategy for the Timed Patrolling

A great number of techniques were already applied to the non-adversarial variation of the multiagent patrolling problem. Experiments suggest that, for general graphs, all those approaches are inferior to a strategy based on the travelling salesman problem (TSP), in which agents are distributed equidistantly along the TSP-cycle. This approach, however, is neither optimal nor scalable. In this article, we present a novel patrolling strategy which tries to overcome these limitations. Inspired on Newton’s law of gravitation, our approach consists in assigning to each node of the graph an abstract mass which grows while the node remains unvisited, creating a force that attracts agents to it. We defined some variations of our approach and experimentally compared them to the TSP-based strategy, concluding that our approach performs better in general.

Towards a context ontology for geospatial data integration

Recently, Geospatial data and Geographic Information Systems (GIS) have been increasingly used As a result, the integration of geospatial data has become a crucial task for decision makers Since GIS and geospatial databases are designed by different organizations using different representation models and there are diverse levels of detail for the spatial features, it is much more complex to achieve data integration in geospatial databases To help matters, context information may be employed to improve two fundamental aspects in Geospatial Data Integration: (1) schema mapping generation and (2) query answering However, a relevant issue when using context is how to better represent context information Ontologies are an interesting approach to represent context, since they enable sharing and reusability and help reasoning In this paper, we propose a context ontology to formally represent context in geospatial data integration We also present an example where this context ontology is used to improve query processing.

A process for the design of Context-Sensitive Systems

Context-Sensitive Systems (CSS) demand that designers consider new aspects and challenges in comparison to traditional applications. In a preliminary experiment, we observed that developers have difficulties when including the concept of context in their applications. This paper presents a CSS Design Process, which details the main activities for context specification and the design of CSS and provides a systematic way to execute these tasks. Two studies were conducted to evaluate the process usage: a detailed execution of each defined activity on the design of a context-sensitive Expert Recommender System; and an evaluation of the process usage by distinct designers on their CSS projects. These studies showed a good acceptance of our proposal and indicated the feasibility of using the process on real projects.

Analyzing On-Line Collaborative Dialogues: The OXEnTCHÊ–Chat

Internet-based virtual learning environments allow participants to refine their knowledge by interacting with their peers. Besides, they offer ways to escape from the isolation seen in the CAI and ITS systems. However, simply allowing participants to interact is not enough to eliminate the isolation feeling and to motivate students. Recent research in Computer Supported Collaborative Learning has been investigating ways to minor the above problems. This paper presents the OXEnTCHE–Chat, a chat tool coupled with an automatic dialogue classifier which analyses on-line interaction and provides just-in-time feedback to both instructors and learners. Feedback is provided through reports, which can be user-specific or about the whole dialogue. The tool also counts on a chatterbot, which plays the role of an automatic coordinator. The implemented prototype of OXEnTCHE–Chat has been evaluated and the obtained results are very satisfactory.

SmartChat – An Intelligent Environment for Collaborative Discussions

Using Computer Supported Collaborative Learning Environments (CSCLE) two or more participants can build their knowledge together, through reflection, collaborative problem resolution, information exchange, and decision-making. The majority of these environments provide tools for communication (e-mails, chats, and forums). However, there are no mechanisms for the evaluation of the interaction contents. The lack of a mechanism to evaluate the interactions could prevent the users from discussing about a specific theme or collaborating among themselves.