eunite

research

applications

knowledge transfer

e-archives



Reflective knowledge models to support an advanced HCI for decision management

Submitted by Josefa Z. Hernandez 27.11.2001, IBA B

Problem

A good communication between human and computer is essential in a lot of computer systems. This issue is especially relevant in real-time management scenarios where the system is intended to provide an intelligent support to human operators in the decision-making task. This type of decision support systems should be capable of warning the operators of any undesired evolution of the installation and recommending adequate control decision that maximise the efficiency of the installation behaviour and minimise the negative impact of the observed problems. As the operators have to assume the system recommendation as their own, it is necessary that the decision support system provides a communicative platform close enough the information needs of the operators that explains its proposals using reasoning schemes and concepts operator are familiar with.

Solution

The Intelligent System Group (ISYS), at the Technical University of Madrid, proposed a new concept of intelligent interface. This intelligent interface is supported by a knowledge-based architecture joined to a user-system dialogue-based interaction model. In a intelligent interface, the way of generating the answer to the userís question responds not only to the criteria regarding the availability of knowledge necessary to reason but also to aspects related to the way of presenting these answers according to characteristics of the user and the interaction context. The central element in the design is the specification of the conversation model to support the user-system communication, that may require the definition of the particular instances of the questions considered according to the characteristics of the problem domain and the requirements of the potential users.

To support this conversational model it has been developed a general architecture with four main components are distinguished: a problem solving medium, composed of a structured collection of basic problem solvers responsible of producing the corresponding answers, a conversation manager, whose goal is to decide the appropriate way to solve problems according to the needs of the conversation, a memory manager, responsible of keeping a consistent set of system beliefs through truth maintenance mechanisms and a presentation manager specialised in the input-output activities of the system.

Status and results

The approach was developed for an application in the domain of private traffic for the city of Turin. The tool for private traffic management may play the role of an automatic assistant for supervising the state of the traffic network and the operation of the units responsible for automatically and dynamically regulating the traffic lights in the intersections of the traffic network. This application was tested, showing the feasibility of the intelligent interfaces approach to support highly interactive decision support system.

Adaptivity and portability

Only adding new problem solvers, if its necessary, and changing the conversational model, the system can be used for other application of control decision support in traffic management field.

More information

J.Z Hernandez, J.M. Serrano. Expert Systems with Applications 19 (2000) 289-304.

technical reports & case studies

executive summaries

scientific papers

tutorials

bibliography

e-dictionary

search eunite with

google



  home