The optimality, complexity and scalability of algorithms are key to real-time communications and practical data collection systems. Algorithms here can include real-time processing of signals in huge quantities, providing quality-of-services to thousands of traffic flows, selecting from hundreds of users to maximize throughput, routing and processing big data in data centres or cloud, or compressing immense sensory data to minimize communication cost. Optimal designs of algorithms are also crucial to cope with non-deterministic elements prevailing in practical network environments, such as non-stationary traffic arrivals, mobility of devices, time-varying wireless channels, and non-persistent energy sources. Algorithms, as well as their optimality, complexity and scalability, are increasingly crucial, as networks and systems scale.
Our expertise in algorithmic designs includes optimization, stochastic control, game theory, Bayesian learning, Markov modelling, and queueing theory. It has been developed to achieve the optimality, reduce the complexity and enhance the scalability of algorithms in a variety of networks and applications. Particularly, we are quoted by IEEE ComSoc to have “succeeded in fulfilling the potential gain of multiuser MIMO for mobile video transmissions”. We have reduced the complexity of radio resource allocation in multiuser wireless channels from NP-hard to linear complexity. We have been able to guarantee the delay of data transmissions, even in the case that the transmission relies on non-persistent renewable energy. We have also decentralized and automated power control in large dense networks to balance interference and stabilize networks.