Mourad Rabah

A Situation-Based Multi-Agent Architecture for Handling Misunderstandings in Interactions

Abstract During interactions, system actors may face up misunderstandings when their local states contain inconsistent data about the same fact. Misunderstandings in interactions are likely to reduce interactivity performances (deviation or deadlock) or even affect overall system behavior. In this paper, we characterize misunderstandings in interactions between system actors (that may be human users or system agents) in interactive adaptive systems. To deal with such misunderstandings and ensure state consistency, we present an agent-based architecture and a scenario structuring approach. The system includes several agents devoted to scenario unfolding, plot adaptation and consistency management. Scenario structuring is based on the notion of a situation that is an elementary building block dividing the interactions between systems’ actors into contextual scenes. This pattern supports not only scenario execution but consistency management as well. In order to organize and control interactions, the situation contextualizes interactions and activity of the system’s actors. It also includes prevention and tolerance agent-based mechanisms to deal with the misunderstandings and their causes. We validate our consistency management mechanisms using Uppaal simulation and provide some experimental results to show the effectiveness of our approach on an online distance learning case study

Trace-based decision making in interactive application: Case of Tamagotchi systems

We present our exploratory work for situation preselecting in interactive applications, assuming that the application is an Interactive Adaptive System based on a sequence of contextualized “situations”. Each situation confines activities and interactions related to a common context, resources and system actors. When one situation is completed, the system has to determine which is the best following one. We introduce in this paper a new preselecting method that identifies possible next situations among all available situations. We propose a strategy using Naïve Bayes based on the analysis of the sets of available traces (the past of users). Combining all obtained results, we get a set of situations, called set of alternatives that can be used in any decision algorithm. We demonstrate our approach on a case study based on Tamagotchi game.

Trace-Based Multi- Cristeria Preselection Approach for Decision Making in Interactive Applications like Video Games

The decision-making in games is essential to make them more automated and smart. A decision algorithm performs its calculations on the set of all the possible solutions. This increases the computation time and may become a combinatorial explosion problem if we have a huge solution space. To overcome this problem, we present our work on relevant solutions preselection before making a decision. We propose a two-steps strategy: i) the first step analyses the system’s traces (users past executions) to identify all the potential solutions; ii) the second step aims to estimate the relevance, called utility, of each of these potential solutions. We get a set of alternative solutions that can be used as an input to any decision algorithm. We illustrate our approach on the Tamagotchi game.

Approche de présélection multicritère à base de traces pour la prise de décision dans les applications interactives de type jeux

La prise de decision dans les jeux est une fonctionnalite indispensable pour automatiser les jeux et les rendre plus autonomes et plus intelligents. Un algorithme de decision effectue les calculs d’optimisation sur l’ensemble des solutions possibles. Cela fait augmenter le temps de calcul et pose un probleme d’explosion combinatoire lorsque l’espace de solutions est grand. Afin de surmonter ce probleme, nous presentons une approche de preselection des solutions pertinentes avant d’effectuer une decision. Notre approche comporte deux etapes : i) utilisation des traces (executions anterieures des utilisateurs) pour identifier toutes les solutions potentielles ; ii) estimation de la pertinence, appelee utilite, de chacune de ces solutions potentielles. Nous obtenons un ensemble de solutions candidates preselectionnees qui sera utilise comme entree a la prise de decision. Nous experimentons notre approche sur un prototype du jeu Tamagotchi.