Ahlem Ben Hassine

Agent-Based Approach to Dynamic Meeting Scheduling Problems

Multi-Agent systems are being more and more widely used to address many distributed combinatorial real-world problems. One such problem is meeting scheduling (MS) that is characterized essentially by two features defined from both its inherently distributed and dynamic nature. In addition, in real world applications, users usually have conflicting preferences, which make the search for an optimal solution an NP-hard problem. However, the majority of the existing works on MS tackle it as a static problem, allow for the relaxation of any constraints and do not deal with achieving any level of consistency. In an attempt to overcome these limitations, the main contribution of this work is a new distributed approach based on the DRAC model (distributed reinforcement of arc consistency) to solve dynamic MS problems. In this approach we authorize only the relaxation of usersý preferences while maintaining arc-consistency on the problem. The underlying protocol is able to efficiently reach optimal solution (satisfying some predefined optimality criteria) whenever possible, using only localized asynchronous communications. This purpose is achieved with minimal message passing and without compromising the privacy of involved users. A comparative analysis divulges that our approach is scalable and worthwhile especially handling strong constraints.

A new distributed approach to solve meeting scheduling problems

Meeting Scheduling (MS) is an important real world-problem that is defined by the process of scheduling meetings while taking into account several constraints related to the availability and preferences of the users. The main contribution of this paper is to propose a new approach based on the DRAC model (distributed reinforcement of arc consistency) to solve the MS problem. A simplified version of this approach was implemented and compared with an approach of the state of the art. The experimental results show that our approach is scalable.

Scheduling meetings with distributed local consistency reinforcement

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.

Horizontal Service Composition for Language Services

In the Language Grid, automatically composing Web services is a crucial task. This task involves vertical and horizontal composition. Vertical composition consists of defining an appropriate combination of simple processes to perform a composition task. Horizontal composition consists of determining the most appropriate Web service from among a set of functionally equivalent ones for each component process. The latter is important in language services. For the horizontal composition of Web services, we propose a generic formalization of any Web service composition problem based on a constraint optimization problem (COP) and then propose an incremental user-intervention-based protocol to find the optimal composite Web service according to some predefined criteria at run-time.

New distributed filtering-consistency approach to general networks

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.

Performing more than AC for hard distributed constraint satisfaction problems

Constraint satisfaction problem (CSP) framework is essentially characterized by the ubiquitous use of local consistency properties and enforcing techniques. The main objective of these techniques is to prune the search space and consequently to enhance the efficiency of the constraint solver. Several levels of local consistency were proposed in the literature among which arc consistency (AC) is the most used one due to its low time and space complexities. However, with the omnipresence of natural distributed real world applications, recently few efforts were directed to enforce local consistency in an entirely distributed manner. Nevertheless, most of these works are limited only to AC property due to the effective-cost of the other existing stronger forms. For some hard constraint network (CN) applying only AC enforcement may be fruitless, case of problems initially arc-consistent. In an attempt to overcome these limitations, the main contribution of this paper is to propose a refinement of the DRAC approach (distributed reinforcement of arc-consistency) to achieve higher level of local consistency, the restricted path consistency (RPC) in a distributed manner with the minimal amount of additional constraint checks, a comprehensive empirical study was performed to highlight the benefit of using the collected knowledge for enforcing arc-consistency on any binary CN, especially for hard arc-consistent problems. The inferred knowledge is used to prune some path inconsistent values (not all of them). The new approach DRAC/sup ++/ is discussed in terms of termination and complexity.