Research On The Information Dissemination Mechanism Among Agents

In the field of distributed artificial intelligence,one agent has been unable to solve the increasingly complex problems,which require multi agents to work together,then multi-agent collaborative work has become a hot research in this field,one of the key questions is how to establish a reliable and effective information dissemination mechanism to rapidly and accurately disseminate the information among agents.However,in the traditional information dissemination mechanism,each agent will select partial channels for information interaction according to the channel states which are monitored spontaneously,this method results in the waste of channel resources;At the same time,due to the limited communication range,it is often necessary to look for a suitable path to send out the information through multi-path forwarding,which makes a great challenge in the process of information dissemination,therefore,it's an urgent need to conduct an in-deep study of multi-channel selection and routing forwarding.The information dissemination among agents is analyzed in the paper,and the main research contents are as follows:(1)Due to the uncertainty of the channel states' perception,a discrete-time Markov channel model is established and the common prediction methods are analyzed,then the HMM is used to predict the channel states through parameter training,concealed state decoding and joint probability solution.(2)The predicted states are formed into a channel state list as the basis of channel selection,and a two-process channel switching mechanism is established,which includes retransmission on the current channel and channel selection and switching based on an adaptive window adjustment strategy,so as to disseminate information on different channels.(3)The basic requirements of routing protocol in multi-agent communication network are given,through the analysis of ant colony algorithm,the superiority of its application in routing protocol is pointed out and the design idea of routing protocol based on ant colony algorithm is given.(4)For the problems in the ant colony algorithm,the path selection and pheromone updating strategy is combined to improve the algorithm,and the improved algorithm is used to design the routing protocol in detail,including message design,route discovery and update,route maintenance,then the routing protocol is set up which can satisfy therequirements of multi-agent communication network.Finally,the main modules of agents are designed and the simulation environment is established,then experiments on multi-channel selection and routing forwarding are carried out respectively,and the main performance indicators are analyzed.The results show that the effectiveness of information dissemination is improved.