Bits and flops in non-ergodic MIMO: Can (a reasonable number of) flops provably offer ergodic-like behavior

Elia, Petros
Habilitation à diriger des recherches (French qualification for conducting Phd theses)

The talk will describe the solution to a sequence of open problems including the problem of establishing meaningful fundamental performance-complexity limits in general MIMO, cooperative and multi-user outage-limited communications. Up to recently the tradeoff between performance and complexity was entirely unknown, and the open problem sparked world-wide interest in inventing more efficient encoders-decoders that come close to the information theoretic limits. The presented work has settled some of these open problems by first succinctly describing the high-SNR fundamental limit between rate, reliability and computational complexity (optimized over all transceivers and all complexity regulating policies), and then by constructively meeting these limits by inventing fast, reliable and efficient MIMO and cooperative encoders, decoders, and protocols that optimally tradeoff performance with complexity.

The talk will also introduce several open problems that relate to the following questions:

What is the complexity price to pay for near-optimal implementation of MIMO, multiuser and cooperative communications?

How does feedback reduce complexity?

How do complexity-constraints affect reliability in different MIMO settings?

How big of a MIMO system (how many transmit antennas or tones or relays or mac users) can your DSP chip sustain?

What is the role of antenna and relay selection in reducing complexity?

What are the cooperative protocols that perform best in the presence of computational constraints?





Systèmes de Communication
Eurecom Ref:
© EURECOM. Personal use of this material is permitted. The definitive version of this paper was published in Habilitation à diriger des recherches (French qualification for conducting Phd theses) and is available at :
See also: