Event

PhD Defense: Joint Scheduling and Precoding in Wireless Networks: A DC Programming Approach

  • Conférencier  Ashok Bandi

  • Lieu

    LU

Please click on this link to connect to the online PhD defense. You should register via the link to get access to the online PhD defense.

Please be informed that Webex events are not accessible on Linux OS.

Members of the defense committee:

  • Assistant Prof. Dr Djamila Aouada, University of Luxembourg, Chairman
  • Prof. Dr Björn Ottersten, University of Luxembourg, Deputy Chairman
  • Dr Bhavani Mysore Shankar University of Luxembourg, SupervisorProf. Dr. Mats Bengtsson, Royal Institute of Technology (KTH),
  • Stockholm, Sweden, Member
  • Dr Joël Grotz, Société Européenne des Satellites (SES), Luxembourg, Member

Unprecedented proliferation of mobile devices is leading to the ever-increasing densification of the networks. Densification and increase in demand overwhelm the dedicated allocation of spatial and frequency resources leading to their reuse. Several techniques, e.g., precoding, have been considered mitigation of interference from this reuse. Despite this, due to resource limitations, only a few requests can be handled during transmission in a given transmit slot. This naturally leads to the scheduling of the requests (or typically referred to as users) which, in turn, impacts co-channel interference. Thus, the average performance of the multiuser multiple-input single-output (MISO) downlink channel, with a large number of users compared to transmit antennas of the base station, depends on the interference management which necessitates the joint design of scheduling and precoding. Typically, the joint design of scheduling and precoding is addressed by mixed-Boolean non-linear programming (MINLP) where the Boolean variables are associated with scheduling. These MINLP formulations in the literature lead to the coupling of Boolean and precoding variables; this coupled nature thwarts the joint update of scheduling and precoding variables. Unlike the previous works which do not offer a truly joint design, this thesis focuses on a joint design that truly facilitates the joint update of scheduling and precoding variable for the following cases in single-cell multiuser MISO downlink channels: per slot joint design for the unicast scenario, per slot design for multigroup multicast (MGMC) scenario and joint design over multiple slots for a unicast scenario. Different challenges in terms of the problem formulation and subsequent reformulations for different metrics in the aforementioned scenarios are discussed.  Through, useful and novel problem formulations combined reformulations, the problems are rendered as difference-of-convex (DC) problems that facilitate the joint update of scheduling and precoding variables. Finally, different algorithms, each focusing on optimizing the corresponding metric, are proposed and their performance is evaluated through numerical results.