Application Of Physarum Model In Network System Construction And Reconfiguration

Web-based environment is usually dynamic, open..., network systems in such an environment need to change to meet the demand for their functionality and quality. Because of a variety of unknown behavior in run-time, network systems with rigid fixed design may cause the failure or untrusted operating results. In that condition, such a system not only needs an optimized design, but also have to change their configuration at runtime-that is, the ability of adaptive reconfiguration. A new research found that a slime mold called Physarum polycephalum forming a network that is both efficient and has high fault tolerance while it foraging. It takes appropriate reconfiguration strategies when encountering unexpected situations in the environment. We believe that adaptive optimization and reconfiguration mechanism of Physarum foraging network can take advantage in a network system. Network systems based on the Physarum foraging rules will both have the ability of optimization, adaptive and reconfiguration.The core mechanism of Physarum forming an efficient way connecting food sources was found by researchers and bind to the Hagen-Poiseuille and Kirchhoff law. The mathematical model is based on two characteristics of Physarum:the adaptive shortest path finding and adaptive network formation, so the model is divided into two main parts:the maze-solving model to get shortest path between two points and multiply sources model for designing efficient network. Corresponding to the two basic models, this dissertation will introduce two different approaches which are used in the following two applications respectively:1, web service composition selection and reconfiguration issues;2, wireless sensor network topology control and reconfiguration.For application1, in this dissertation, based on maze model, a web service composition selection and reconstruction method is presented. Combinations choices are modeling to a connected graph, web service components corresponding to nodes in the graph and their service qualities as edge weights. Shortest path selection method of Physarum corresponding to the optimal quality of Web services portfolio selection, our method can not only get the optimal combination of services but also be able to do adaptive reconfiguration change in the condition that service node or network failure in order to achieve the best.For application2, dissertation first show the similarity of the network of wireless sensor network topology and Physarum multi-sources network using graph as a media, and then analyzes the characteristics and needs of the wireless sensor network topology control, a bio-inspired wireless sensor network topology control and reconfiguration strategy is presented. With a definition of relation neighbors, the bionic method combining with neighboring plan can solve the topology control and reconfiguration problem.At the end, experimental comparative validation and visualization analyze verify the feasibility and effectiveness of the proposed bionic methods.