A Strategy Of Location Management Based On Dynamic Probability In Mobile Environment

With the fast development of mobile telecommunication, the computing environment has become so different with the one used before. The widely usage of the mobile terminals is such a fact that the research of the applications in mobile environment, namely mobile computing, has been a hot field. And location management has become one of the most important issues in the field of mobile computing.With a main job to locate mobile unit in an efficient way, the location management is a vital part for most of applications in mobile environment. The location management in mobile environment is a typical problem of trade-off. An efficient strategy of location needs to make balance between updating and paging mobile units'locations. The research of location management mainly focuses on reducing the total cost including the cost of location updating and location paging.The strategy of location management based on dynamic probability, namely DPB, maintains a K-layer possible resident region (PRR-K) for each mobile unit by using the probability computing models and the technologies of location prediction and data mining. With the mobility patterns of mobile units, DPB uses the technologies of data mining to gain the probability information of the resident probabilities and direction probabilities for every cell. According to the changes of the movement state of mobile units, DPB uses the probability information to compute the related probabilities for mobile units (MU). When the server needs to call a mobile unit, the PRR-K information will be fetched from the Base Station which maintains the related information for this mobile unit. With this probability information, the server will be able to launch a paging process by dividing the location area into several paging zones.As another form of DPB, CPDB only uses the direction probabilities to make the prediction of the location of MU, so it can reduce the cost of maintaining the location information. However, as the direction probabilities are gained by using the technology of data mining on the historical datum, CDPB will also be able to have a good performance on location management.