In this paper, we introduce open cellular learning automata and then study its convergence behavior. It is shown that for a class of rules called commutative rules, the open cellular learning automata in stationary external environments converges to a stable and compatible configuration. The numerical results also confirm the theory.  

A cellular learning automaton (CLA) is a collection of learning automata arranged in a ‎grid similar to cellular automata and interacts with each other. Each learning automaton ‎based on the actions chosen by their neighbors tries to find its best action in order for the ‎cellular learning automata to reach a particular goal. In this paper several applications of ‎cellular learning automata to designing image processing operations such as noise ‎removal image segmentation and feature extraction are presented. The proposed ‎algorithms have number of good characteristics such as: effectiveness in the presence of ‎noise higher accuracy .comparing to other image processing algorithms parallel nature ‎and locality.‎

In this paper, a new algorithm which is the result of combination of cellular learning automata (CLA) and shuffled frog leap algorithm (SFLA) is proposed for optimization of functions in continuous, static environments. In the frog leaping algorithm, every frog represents a feasible solution within the problem space. In the proposed algorithm, each memeplex of frogs is placed in a cell of CLA. learning automata in each cell acts as the brain of memeplex and will determine the strategy of motion and search. The proposed algorithm along with the standard SFLA and two global and local versions of particle swarm optimization algorithm have been tested in 30-dimensional space on five standard merit functions. Experimental results show that the proposed algorithm has a performance of the introduced algorithm is due to the control of search behavior of frogs during the optimization process.

Variety of methods for traffic policing such as windowing algorithm and leaky bucket for computer networks have been proposed in the literatures. The windowing algorithm works well for packet networks but it is not appropriate for High speed networks such as ATM. In this paper, a new group of traffic policers for ATM network has been proposed. These proposed traffic policers are obtained by improving the leaky bucket method. learning automata is used to adapt the parameters of the leaky bucket according to the behavior of the traffic sources. To study the effectiveness of the proposed methods, computer simulations have been conducted. The results of simulation have shown that the proposed methods reduce the percentage of the cell loss. The proposed traffic policer works well for the situations where the traffic characteristics. Of the source are unknown.

Active database systems (ADBS) can react to the occurrence of predefined events automatically by definition a collection of active rules. One of the most important modules of ADBS is the rule scheduler, which has considerable impact on performance and efficiency of ADBS. The job of rule scheduler is the selection of a rule for execution from the set of ready for execution rules. In this paper, we propose a new approach based on learning automata to improve the rule scheduling performance in terms of average response time, response time variance, and throughput. learning automata have been used to obtain better estimations for rule execution probabilities. The results of experimentations show that the performance of the proposed method outperforms the most effective existing rule scheduling method.