Towards a fuzzy evaluation of the adaptivity degree of an evolving agent

Proceedings of the 5th International Conference on Soft Computing as Transdisciplinary Science and Technology - CSTST '08

2008

Self Cite

The use of multirobot systems, is affecting our society in a fundamental way; from their use in hazardous environments, to their application in automated environmental cleanup. In an unknown environment, one of the most important problem related to multirobot systems, is to decide how to coordinate actions in order to achieve tasks in an optimal way. Ant algorithms are proved to be very useful in solving such distributed control problems. We introduce in this paper a modified version of the known ant algorithm, called Counter-Ant Algorithm (CAA). Indeed, the robots' collaborative behaviour is based on repulsion instead of attraction to pheromone, which is a chemical matter open to evaporation and representing the core of ants' cooperation. In order to test the performance of our CAA, we implement, simulate and test our algorithm in a generic multirobot environment. In practical terms, the subdivision of the cleaning space is achieved in emergent and evolving way. A series of simulations show the usefulness of our algorithm for adaptive and cooperative cleanup.

Section: Discussionmentioning

confidence: 99%

Counter-ant algorithm for evolving multirobot collaboration

Chatty

Proceedings of the 5th International Conference on Soft Computing as Transdisciplinary Science and Technology - CSTST '08

2008

Self Cite

“…As the old paths marked by the pheromone will disappear, new places become open to explore, which were not detected by other ant-robots. This leads us to say that the subdivision of the area to explore is self organized [9]. We present also a solution for the stagnation recovery which allows the robots to overtake dead ends.…”

Section: Discussionmentioning

confidence: 99%

Self-Organizing Multirobot Exploration through Counter-Ant Algorithm

Chatty

Self-Organizing Systems

2008

Self Cite

Abstract. This paper presents an evolving method for a self-organizing multirobot exploration of an unknown environment. In such problem, a big consideration is given to the coordination behavior of robots in order to achieve the common tasks in an optimal way. Ant algorithms are proved to be very useful in solving such distributed control problems. We present here a modified version of the known ant algorithm, called Counter-Ant Algorithm (CAA). Indeed, the robots'collective behavior is based on repulsion instead of attraction to pheromone, which is a chemical matter open to evaporation and representing the core of ants' cooperation. A series of experimentations with MINDSTORMS LEGO robots, and simulations under Madkit platform, in laboratory conditions similar to real ones, show the usefulness of our algorithm for self-organizing and cooperative exploration.

“…Such systems are proved to be efficient in giving natural and approximate reasoning, i.e. in [6] and [14].…”

Section: Fuzzy Parametersmentioning

confidence: 99%

ARG_FuzzySim: Concluding Failed Multilateral Argumentation-Based Negotiation

Barhoumi

2008 International Conference on Computational Intelligence for Modelling Control &Amp; Automation

2008

Self Cite

In this paper, we propose a new model that gives a solution about the conflict of interests in a multiagent system, and thus to conclude a failed argumentation based negotiation (ABN). Baptized ARG_FuzzySim, our model starts from the past exchanged arguments and integrates the concept of fuzzy similarity to determine the most similar argument and subsequently, the final decision. Among the strong features of ARG_FuzzySim is its facility to implement and propose a solution for conflicting argumentative negotiation between two or several agents. Various applications based on different negotiation scenario show the practicability and the effectiveness of the suggested model.