Objective of this course is to introduce the student to advanced techniques for medium access control, coordinated or contention based, and radio resource management in both cellular systems and mesh or ad-hoc networks. Power allocation, rate adaptation and scheduling will be discussed both in centralized and distributed settings. Some mathematical methods play a fundamental role in resource allocation, namely, classical Perron-Frobenius theory for nonnegative matrices, convex and nonconvex constrained optimization, distributed optimization and game theory. The course introduces the student to such methods and exemplifies their application to various resource allocation problems. Additionally, the course addresses relevant aspects of resource allocation in wireless networks such as fairness and cross-layer design.
Properties and challenges of the wireless medium.
Basic concepts of communication networks: the layered architecture
The 802.11 wireless architecture
Evolution of wireless cellular network architectures: From Global System for Mobile to Advanced-Long Term Evolution
Multiple Access Schemes: CSMA variants, TDMA, FDMA, CDMA, OFDMA, SC-FDMA, SDMA
Uplink-downlink duality
Opportunistic scheduling and multiuser diversity
Advanced concepts: small cells and heterogeneous networks, relaying and cooperation, network coding, cognitive radio networks
Basics of resource allocation: power allocation, rate adaptation, and scheduling
Classical resource allocation techniques: Centralized and distributed power control based on the Perron-Frobenius theorem
Fundamentals of convex constrained optimization and application to resource allocation
Resource allocation and fairness
Fundamentals of nonconvex optimization and relaxation techniques
Applications of nonconvex optimization to resource allocation
Fundamentals of distributed optimization and applications to resource allocation
Fundamental concepts of game theory
Resource contention via game theoretical methods
Resource contention and random access protocols
Concepts of Discrete Time Markov Chains (DTMC)
Design and performance analysis of random access protocols via DTMC
Describes and/or recognizes wireless channel models
Compares different CSMA schemes in various communication media and explains how to combat the hidden node effect in wireless systems
Criticizes the limits of the a layered architecture in wireless systems
Defends the use of cross-layer design in wireless network
Applies rate adaptation schemes to maximize the throughput in IEEE 802.11
Compares handover schemes of different cellular architectures
Appraises and compares the distribution of functionalities in network entities for different architectures
Argues on the pros and contras of different multiple access schemes according to various criteria (e.g. spectral efficiency, power efficiency, robustness to interference)
Compares and contrasts micro-diversity and various macro-diversity schemes
Computes the total rate of SDMA with various receivers
Relates the multiple access in uplink to broadcasting in downlink and justifies the concept of uplink-downlink duality
Uses uplink-downlink duality to design a precoder and allocate power
Contrasts multiple access in uplink and broadcasting in downlink in terms of channel state acquisition both for TDD and FDD transmission
Uses multiuser diversity for opportunistic scheduling
Compares multiuser diversity for users having identical and different channel statistics
Contrasts opportunistic scheduling in terms of channel state acquisition and feedback both for uplink and downlink and for both FDD and TDD transmission schemes
Appraises the impact of multiple antennas on opportunistic scheduling
Analyses different settings with interference in small cells and designs countermeasures
Categorizes relaying schemes in LTE
Analyses performance of relaying schemes
Argues on possible improvements of relaying schemes via network coding and physical layer network coding
Uses the Perron-Frobenious theorem to allocate power in a centralized manner
Judges the feasibility of a power control problems and formulates alternative approaches in case of unfeasibility
Uses the Perron-Frobenious theorem to design a distributed power control scheme
Judges the convergences of distributed power control based on the Perron-Frobenius theorem and appraises the robustness of asynchronous power control
Applies techniques of convex optimization to discriminate convex problems and determine necessary and/or sufficient conditions for global optimality
Judges the applicability of KKT conditions and duality
Uses KKT conditions to solve convex optimization problems
Uses duality to solve convex optimization problems
Applies convex optimization to resource allocation in wireless communications
Compares different definitions of fairness and applies them to rate allocation
Appraises the effect of channel knowledge at the transmitter on different fairness criteria
Applies KKT conditions for opportunistic user scheduling
Describes a proportional fair algorithm for opportunistic scheduling
Applies relaxation to nonconvex quadratic constrained quadratic programming
Formulates resource allocation problems as constrained optimization programming
Contrasts various distributed optimization methods
Applies the concept of best response to determine Nash equilibria
Argues about existence and uniqueness of Nash equilibria
Assesses if a given game is a potential game and solves it
Defends the concept of Pareto optimality in resource allocation
Contrasts the concepts of pure and mixed strategies in game theory
Uses coupled constrained concave game to allocate powers in heterogeneous networks
Uses discrete time Markov chain to model specific aspects of communication systems
Discusses the meaning of Markov chain transition matrices interpreting Kolmogorov-Chapman equations
Contrasts stationary and limiting distributions of a Markov chain and predicts when they coincide in specific cases
Applies Markov chains to Aloha systems and analyzes a slotted Aloha protocol
Justifies the usefulness of exponential back-off in CSMA via analysis with Markov chain