On-line Data Retrieval Algorithm with Restart Strategy in Wireless Networks

Abstract

Data retrieval is an efficient scheme for data dissemination in mobile computing, and its primary goal is to improve the waiting time and save the energy of mobile devices. Existing data retrieval algorithms are developed at the basis of a-priori knowledge of wireless data broadcast. In practical, they are not suitable for on-line data retrieval where the knowledge is not known. In this paper, we study on-line data retrieval problem with Las Vegas strategy of randomized algorithm, called as LVDR, and reduce the competitive rate of this problem, , where is the number of channels and is the length of broadcast cycle. Thus, we observe that different schemes of choosing channel will generate the different efficiencies in randomized algorithm. Therefore, we propose another scheme introduced restart strategy to execute the other retrieval sequences of choosing channels when the client may not retrieve the requested data item in certain time, called as LVDR-RS. In addition, we adopt Markov chain to predict the optimal restart cycle for each requested data item. Through analysis, we obverse that the competitive rate is closely related with the number of restarts and restart cycle