Ilhem Kallel

MAGAD-BFS: A learning method for Beta fuzzy systems based on a multi-agent genetic algorithm

This paper proposes a learning method for Beta fuzzy systems (BFS) based on a multiagent genetic algorithm. This method, called Multi-Agent Genetic Algorithm for the Design of BFS has two advantages. First, thanks to genetic algorithms (GA) efficiency, it allows to design a suitable and precise model for BFS. Second, it improves the GA convergence by reducing rule complexity thanks to the distributed implementation by multi-agent approach. Dynamic agents interact to provide an optimal solution in order to obtain the best BFS reaching the balance interpretability-precision. The performance of the method is tested on a simulated example.

An adaptive vehicle guidance system instigated from ant colony behavior

In view of the high dynamicity of traffic flow and the polynomial increase in the number of vehicles on road networks, the route choice problem becomes more complex. A classical shortest path algorithm based only on road length is no longer relevant. We propose in this paper an adaptive vehicle guidance system instigated from the ants behavior, well known for its good adaptativity; this system allows adjusting intelligently and promptly the route choice according to the real-time changes in the road network situations, such as new congestions and jams. This method is implemented as a deliberative module of a vehicle ant agent in a collaborative multiagent system representing the entire road network. Series of simulations, under a multiagent platform, allow us to discuss the improvement of the global road traffic quality in terms of time, fluidity, and adaptativity.

Adaptation capability of cognitive map improves behaviors of social robots

In this paper, we study the impact of the cognitive maps adaptation in the context of multi-robot system. This map governs the emergence of non-trivial behaviors and structures at both individual and social levels. In particular, we show that adding a simple imitation and deposit behavior allows the cognitive robots to adapt themselves in unknown environment to solve different navigation tasks. We show that in our architecture the individual discoveries in each robot (i.e., goals) can have an effect at the population level, which induce then a new learning at the individual level and reciprocally, from the individual to the population level. We performed a series of experimentations with robots and simulated agents to validate our system.

Emergent complex behaviors for swarm robotic systems by local rules

This paper describes a clustering process taking inspiration from the cemetery organization of ants. The goal of this paper is (i) to show the importance of the local interactions which allow to produces complex and emergent behavior. (ii) To propose a multi-robot systems in the field of clustering objects allowing optimization of: time of convergence, rate of occupation of the objects in the environment and final number of the clusters. And finally (iii) to propose another system with the use of cognitive robots instead of reactive robots in the same field of clustering objects with generic rules which can be used independently of the environment. Series of simulations enable us to discuss and validate the proposed approach.

MARCoPlan: MultiAgent Remote Control for Robot Motion Planning

A multiagent system to support a mobile robot motion planning has been presented. Baptized MARCoPlan(MutiAgent Remote Control motion Planning), this system deals with optimizing robot path. Considered as an agent, the robot has to optimize its motion from a start position to a final goal in a dynamic and unknown environment, on the one hand by the introduction of sub-goals, and on the other hand by the cooperation of multiagents. In fact, we propose to agentify the proximity environment (zones) of the robot; cooperation between theses zones agents will allow the selection of the best sub-goal to be reached. Therefore, the task of the planner agent to guide the robot to its destination in an optimized way will be easier. MARCoPlanis simulated and tested using randomly and dynamically generated problem instances with different distributions of obstacles. The tests verify some robustness of MARCoPlanwith regard to environment changes. Moreover, the results highlight that the agentification and the cooperation improve the choice of the best path to the sub goals, then to the final goal.

Improvement of the road traffic management by an ant-hierarchical fuzzy system

In view of dynamicity on road networks and the sharp increase of traffic jam states, the road traffic management becomes more complex. It is clear that the shortest path algorithm based only on road length is no longer relevant.We propose in this paper a hybrid method based on two stages based on ant colony behavior and hierarchical fuzzy system. This method allows adjusting intelligently and promptly the road traffic according to the real-time changes in the road network states by the integration of an adaptive vehicle guidance system. The proposed method is implemented as a deliberative module of a vehicle ant agent in a collaborative multiagent system representing the entire road network. Series of simulations, under a multiagent platform, allow us to discuss the improvement of the global road traffic quality in terms of time, fluidity, and adaptability.

Ranking criteria based on fuzzy ANP for assessing E-commerce web sites

Assessing E-commerce web sites quality is essential not only to have recommendations for improvement but also to make comparisons with competitors. In this paper, the aim is to know the best criteria for the evaluation and obtain a weight for them using fuzzy Analytic Network Process (fuzzy ANP). The subjective judgments of the decision maker are expressed by fuzzy numbers. The decision making problem is solved by making fuzzy pairwise comparisons and a feedback between the criteria.

The effect of learning by imitation on a multi-robot system based on the coupling of low-level imitation strategy and online learning for cognitive map building

It is assumed that future robots must coexist with human beings and behave as their companions. Consequently, the complexities of their tasks would increase. To cope with these complexities, scientists are inclined to adopt the anatomical functions of the brain for the mapping and the navigation in the field of robotics. While admitting the continuous works in improving the brain models and the cognitive mapping for robots’ navigation, we show, in this paper, that learning by imitation leads to a positive effect not only in human behavior but also in the behavior of a multi-robot system. We present the interest of low-level imitation strategy at individual and social levels in the case of robots. Particularly, we show that adding a simple imitation capability to the brain model for building a cognitive map improves the ability of individual cognitive map building and boosts sharing information in an unknown environment. Taking into account the notion of imitative behavior, we also show that the individual discoveries (i.e. goals) could have an effect at the social level and therefore inducing the learning of new behaviors at the individual level. To analyze and validate our hypothesis, a series of experiments has been performed with and without a low-level imitation strategy in the multi-robot system. Graphical Abstract

Hybrid Fuzzy-MutiAgent planning for robust mobile robot motion

This paper presents an intelligent hybrid system to support the planning for a mobile robot motion in unknown and dynamic environment. Called Fuzzy-MARCoPlan (Fuzzy-MultiAgent Remote Control motion Planning), this system optimizes the path by the introduction of sub-goals and through a multiagent cooperation based on fuzzy reasoning. In fact, we propose to agentify the surrounding zones of the robot; these zone agents compete for attracting the sub-goal. A planning agent, fortified with a fuzzy rule based system, decides on the best sub-goal to reach. Fuzzy-MARCoPlan is simulated and tested on several navigation environments which are generated randomly under the multiagent platform MadKit. These tests confirm the robustness of the proposed system in terms of path optimality in a dynamic environment. Moreover, the obtained results reinforce the advantage of a multiagent planning hybridized with fuzzy reasoning for mobile robot motion planning.

Assessing web sites quality: A systematic literature review by text and association rules mining

Assessing web sites is considered as a Multiple Criteria Decision Making problem (MCDM), with a massive number of criteria; a reduction phase is needed.Text mining is applied for this SLR to construct a dataset of criteria.Association Rules Mining are used to study interdependencies between criteria and the category of the web site. Nowadays society is deeply affected by web content. A web site, regardless of its category, can provide or not for users their needs. To identify its strengths and weaknesses, a process of analyzing and assessing its quality, via some criteria, is necessary. Assessing web sites is considered as a Multiple Criteria Decision Making problem (MCDM), with a massive number of criteria; a reduction phase is needed. This paper presents, firstly a Systematic Literature Review (SLR) to identify the purposes of recent researches from the assessment and determine the affected categories; secondly, it proposes a process of collecting and extracting data (criteria featuring web sites) from a list of studies. Text mining is applied for this SLR to construct a dataset. Then, a method based on Apriori algorithm is assigned and implemented to find association rules between criteria and the category of the web site, and to get a set of frequent criteria. This paper also presents a review on soft computing assessing methods. It aims to help the research community to have a scope in existing research and to derive future developments. The obtained results motivate us to further probe datasets and association rule mining.