An increasing on knowledge of MAS trained by Boltzmann machine algorithm based sugarscape CA using a synergy of communication and cooperation bet agents

Sugarscape model is a multi-agent environment that is used for modeling and organizing processes such as social, political and economic. After the previous studies which were concerned with the production of a learned multi-agent model based on Boltzmann Machine learning algorithm and also the evaluation of the learning of a learned system in sugarscape, the purpose of this study is to evaluate the learning done after adding the two parameters of communication and cooperation to the sugarscape learned model. Thus a cellular learned multi-agent model with use of Boltzmann Machine learning algorithm based on sugarscape model was considered. In this model, each agent has been allocated with a parameter that indicates the knowledge of the agent. Once all agents reach sugar peaks it means that all agents have become knowledgeable and the model has converged. After that the two parameters of communication and cooperation are added to the given model and for each one of the models the number of agents present in sugar peaks after the model had reached convergence per the specific number of agents has been measured. After analyzing the resulting diagram it was concluded that after the convergence of the model, the average number of knowledgeable agents in learned model with communication and cooperation is higher than the number of knowledgeable agents in learned model without use of communication and cooperation. Therefore communication and cooperation of the agents causes to increase the learning done in the environment.