Design And Implementation On A Bio-network Model Based On Mobile Agent

With a desirable requirement from people’s application, Internet is thereby expected towards to applications with high performance, multiplicity and large-scale. Therefore people claim Internet is to be user-oriented, scalability, mobility, security, robustness and adaptivity. Many mechanisms employed by biosystem, such as immune system, evolution theory, pave a new way to explore new type network to accommodate those requirements. In this thesis, we apply the principles and mechanisms of biosystem into network technology to develop a type of bionetwork to balance loading and to evaluate Quality of Service (QoS).We abstract the principle from large-scale biosystem first, and then we design a new bionetwork model by combining the principle with the mobile agent technology and biosystem mechanism, of which includes two parts of bioentity and bionetwork platform. In details, the bioentity comprises its behaviors, expression and implementation, producing multiplicity, and services can be provided, while the bionetwork platform consists of bioentity context, bionetwork service, and bionetwork container. The thesis makes efforts on the design for communication of the bionetwork model, of which consists communication between bioentities, between bioentity and cyber entity, and between cyber entities. With this bionetwork model, we design a scheduling algorithm based on immune genetic to balance network loading, which is designed based on scheduling entity and performed by the scheduling entity through communicating between bioentities. Then we design a QoS evaluation model within the bionetwork model frame, especially we detail its architecture, designing and optimization. The QoS evaluation model is designed based on the fuzzy logical and neuron network, and is optimized with genetic algorithm, therefore, it can evaluate QoS performance under parallel processing in large-scale network by taking the advantages of learning ability and non-linearity of neuron network, and of fuzzy reasoning ability of fuzzy logical.We implement and simulate the scheduling algorithm for network balancing loading, and evaluate the algorithm performance by comparing it with traditional immune algorithm. The simulation results show that the scheduling algorithm designed based on genetic immune algorithm can considerably meet our goal of balancing network loading, and takes on a better performance compared with traditional algorithm, and hereby enhance the self-adaptive ability of network. Furthermore, we simulate the QoS evaluation model in the bionetwork frame, and the results show that the QoS evaluation model can perform commendably within the designed bionetwork frame and provides a high performance.