| With the rapid development of communication technology and computingtechnology, mobile device represented by the smartphone gets rapid development,and mobile applications have sprung up; so smartphone has become an indispensablepart of human life. Compared with the PC, smartphone is still resource-limited,especially the capacity of battery. At present, data transmission and data processingare the main energy consumption ways of the smartphone, and limited batterycapacity and people growing energy demands make the issues on smartphone energybecome more obvious. It becomes the hotspot for researchers to solve the problemabout how to optimize the energy consumption of smartphone from two aspects: thedata transmission and data processing, to ensure daily use for users.Around these problems, this article aims to optimize the device energyconsumption and extend device resources; and it expands from two aspects: lower thedata transmission and energy consumption of data processing. The specific contentsof this paper are as follows:(1) Present the self-adaption multiple order prediction model—Markov underthe cloud environment. According to the WiFi trajectory of the mobile users, analyzethe trajectory of the mobile users to find its inherent law and related access method;propose the self-adaption multiple order prediction model-Markov under the cloudenvironment to predict whether there is a better WiFi network in its future position,with the information of the mobile users’ WiFi global position distribution, whichprovides decision support for the dynamic link handoff approach.(2) Present the dynamic link handoff approach. According to users’ delaytolerance to application and the predicting results of WiFi to choose whether to delaythe data transmission waiting for a better WiFi network. In aspect of the delay taskmanagement, we design the FCFS (First Come First Serve) task dispatch to managerelated delay task; and in the process of data transmission, due to network fluctuation,we design the strategy of the detecting data transmission to maintain mobile devicesconnect with the optimal wireless network to optimize the energy consumption of data transmission.(3) Design a method that can process offline the WiFi trajectory of mobile usersbased on MapReduce. Analyze users’ WiFi trajectory, use the MapReduce’s powerfulability of processing the distributed data and the HBase’s powerful capacity of datastorage, and then calculate the WiFi trajectory of mobile users and build thecorresponding multiple order transition probability matrix of Markov. When aapplication request from user arrives, we can quickly respond to it and return relatedprediction results.(4) Design the static partitioning method of mobile applications under the cloudenvironment. Analyze the function structure of mobile application in detail, andpresent the static partitioning method of mobile application under the cloudenvironment and partition the application by using the rich resources of cloudcomputing. Through wireless network, we transfer high intensive computation andhigh energy consumption subtasks to the cloud server, the smartphones only performsimple subtasks. When tasks are executed, it will return the result to the mobiledevice.In this thesis, the innovative work includes the following:(1) Present the self-adaption multiple order prediction model-Markov under thecloud environment, use Hadoop technical analysis to process user track information,and combining with multiple order prediction model—Markov, predict its futureposition according to user’s WiFi trajectory.(2) Present the dynamic link handoff approach; analyze users’ delaytransmission tolerance to application, and then combining with the prediction resultsof WiFi, make it come true that the mobile device delays to connect the datatransmission from optimal wireless network. |
