رسائل ماجيستير
English
2017
Rate Aware Network Codes For Coordinated Multi Base-Station Networks
Mohammed Hameed Abdallah Saif , Sameh Sorur
كلية الهندسة-جامعة الملك فهد للبترول والمعادن · السعودية
In this thesis work, we study the problem of minimizing the message download completion delay from multiple base stations system. By utilizing a multi-Base station Rate Aware Instantly Decodable Network Coding (MB-RA IDNC) thecompletion time is minimized. IDNC can speed up the download process by taking advantage of the users’ side information. The different rates of the base-stationsto the various users will be thus incorporated in the network coding decisions, so as to schedule the coded messages and transmission rates jointly in order to reduce the overall completion time. This work considers asymmetric rates for users which reflects more practical application scenarios and gives a new trade-off between the selection of message combination for different users and the broadcasting rate to reduce the delivery time of the wanted messages for the users.However, more than one base station will target the same user simultaneously causing transmission conflicts. This model considers the transmission conflicts among the base stations in finding the optimal message combinations and transmission rates. Using this model, we design a multi-layer heuristic algorithm which is used in solving thedownload time minimization problem in such systems. The performance of the proposed algorithm (in terms of the completion delay) with perfect channel is compared to that of the classical IDNC and uncoded schemes in which every base station tries to minimize its own completion delay without considering the rate awareness. Also, a message placement policy in the base stations is simulated and compared to the performance of classical IDNC and uncoded schemes when the messages are distributed between base stations without repetitions. The proposed scheme for perfect channel with message placement policy is shown to achieve significant reduction in completion delay as compared to the classical IDNC and uncoded schemes. Furthermore, a channel erasure probability is investigated (i.e.,when message loss event occurs). Under such scenario, the uncertainty about message reception at the targeted user would drive the base station to carry out the message combinations selection and re-transmit the lost message. Therefore, the performance of multi-layer MB-RA IDNC algorithm is studied and derived for such scenario.