This page highlights a few research areas where we have made distinct contributions. Naturally, the page is selective, both in terms of research topics and the papers listed. For a more complete picture, please see the publications page and the individual home pages of past and present students, postdocs and collaborators listed on the group page.
Machine learning
Our research focuses on techniques that enable faster and more efficient training of machine learning models. We have developed theory and methodology for parallel and asynchronous computations, investigated how traditional algorithms can be tuned, adapted and accelerated to produce accurate results faster, and studied how the information exchange between compute nodes can be reduced to save energy and shorten the computation times. Our interests range from fundamental theory, via algorithm development and implementations, to applications.

Selected publications

Asynchronous iterations in optimization: new sequence results and sharper algorithmic guarantees

H. R. Feyzmahdavian and M. Johansson

Journal of Machine Learning Research, 2023

PDF

A fast and accurate splitting method for optimal transport: analysis and implementation

V. V. Mai, J. Lindbäck and M. Johansson

International Conference on Learning Representations (ICLR), 2022

PDF

Advances in asynchronous parallel and distributed optimization

B.M. Assran, A. Aytekin, H.R. Feyzmahdavian, M. Johansson and M.G. Rabbat

Proceedings of the IEEE, November 2020

PDF

Anderson acceleration of proximal gradient methods

V.V. Mai and M. Johansson

International Conference on Machine Learning (ICML), July 2020

PDF

Convegence of a stochastic gradient method with momentum for nonsmooth nonconvex optimization

V.V. Mai and M. Johansson

International Conference on Machine Learning (ICML), July 2020

PDF

The convergence of sparsified gradient methods

D. Alistarh, T. Hoefler, M. Johansson, S. Khirirat, N. Konstantivov and C. Renggli

Neural Information Processing Systems (NeurIPS), 2018.

PDF

Compressed gradient methods with Hessian-aided error compensation

S. Khirirat, S. Magnússon and M. Johansson

IEEE Transactions on Signal Processing 2020

PDF

A flexible framework for communication-efficient learning: from HPC to IoT

S. Khiriat, S. Magnússon, A. Aytekin and M. Johansson

AAAI Conference on Artificial Intelligence, 2020

PDF

Global convergence of the heavy-ball method for convex optimization

E. Ghadimi, H.R. Feyzmahdavian and M. Johansson

European Control Conference, 2015

PDF

Exploiting serverless runtimes for large-scale optimization

A. Aytekin and M. Johansson

IEEE Cloud, 2019

PDF

POLO: a policy-based optimization library

A. Aytekin, M. Biel and M. Johansson

arXiv preprint, 2018.

PDF

An asynchronous mini-batch algorithm for regularized stochastic optimization

H.R. Feyzmahdavian, A. Aytekin and M. Johansson

IEEE Transactions on Automatic Control, 2016

PDF

A delayed proximal-gradient method with linear convergence rate

H.R. Feyzmahdavian, A. Aytekin and M. Johansson

IEEE MLSP, 2014

PDF

Ergodic mirror descent

J. Duchi, A. Agarwal, M. Johansson and M.I. Jordan

SIAM Journal on Optimization, 2012

PDF
Complex control systems
Automatic control explores how computers can be used to control the physical world in a safe, predictable and efficient manner. Control systems are pervasive. They control the indoor climate in our homes; adjust the power that our cell phones use to communicate reliably; enhance efficiency and reduce emissions in factories and power plants; and improve comfort, safety and performance of cars and airplanes. Our research has focused on advancing the theory and engineering methodology for complex control systems, including hybrid systems, networked control systems and systems that operate asynchronously and with delayed information.

Selected publications

Asymptotic stability and decay rates of homogeneous positive systems with bounded and unbounded delays

H.R. Feyzmahdavian, T. Charalambous and M. Johansson

SIAM Journal on Control and Optimization, 2014

PDF

Exponential stability of homogeneous positive systems of degree one with time-varying delays

H.R. Feyzmahdavian, T. Charalambous and M. Johansson

IEEE Transactions on Automatic Control, 2013.

PDF

Networked control systems

A. Bemporad, M. Heemels and M. Johansson

Springer Verlag, 2010

PDF

On the trade-off between communication and control cost in event-triggered dead-beat control

B. Demirel, V. Gupta, D.E. Quevedo and M. Johansson

IEEE Transactions on Automatic Control, 2016.

PDF

Modular design of jointly optimal controllers and forwarding policies for wireless control

B. Demirel, Z. Zou, P. Soldati and M. Johansson

IEEE Transactions on Automatic Control, 2014.

PDF

Networked estimation under contention-based medium access

M. Rabi, L. Stabellini, A. Proutiere and M. Johansson

International Journal of Robust and Nonlinear Control, 2010.

PDF

Optimal stopping for event-triggered sensing and actuation

M. Rabi, K.H. Johansson and M. Johansson

IEEE Conference on Decision and Control, 2008.

PDF

PID controller tuning rules for varying time-delay systems

L.M. Eriksson and M. Johansson

American Control Conference, 2007

PDF

Joint optimization of communication rates and linear systems

L. Xiao, M. Johansson, H. Hindi, S.P. Boyd and A. Goldsmith

IEEE Transactions on Automatic Control, 2003.

PDF

Piecewise linear control systems: a computational approach

M. Johansson

Sprinver Verlag, 2003

PDF

Piecewise linear-quadratic optimal control

A. Rantzer and M. Johansson

IEEE Transactions on Automatic Control, 2000

PDF

Piecewise quadratic stability of fuzzy systems

M. Johansson, A. Rantzer and K.-E. Årzén

IEEE Transactions on Fuzzy Systems, 1999

PDF

Computation of piecewise quadratic Lyapunov functions for hybrid systems

M. Johansson and A. Rantzer

IEEE Transactions on Automatic Control, 1998.

PDF
Decision support technology
Human decision-making can be aided by computer programs which suggest viable solutions and allow the decision-maker to radidly visualize and explore different options. We have developed decision support systems for several application domains, including machine scheduling, sports, bioinformatics, and control.

Selected publications

Underground mine scheduling of mobile machines using constraint programming and large neighborhood search

M. Åstrand, M. Johansson and A. Zanarini

Computers and Operational Reesearch, 2020

PDF

Scheduling double round-robin tournaments with divisional play using constraint programming

M. Carlsson, M. Johansson and J. Larson

European Journal of Operational Reesearch, 2017

PDF

Ultrasome: efficient aberration caller for copy number studies of ultra-high resolution

B. Nilsson, M. Johansson, F. Al-Shahrour, A.E. Carpenter and B.L. Ebert

Bioinformatics, 2009

PDF

Threshold-free high-power methods for the ontological analysis of genome-wide gene-expression studies

B. Nilsson, P. Håkansson, M. Johansson, S. Nelander and T. Fioretos

Genome Biology, 2007

PDF

Interactive tools for education in automatic control

M. Johansson, M. Gäfvert and K.J. Åström

IEEE Control Systems Magazine, 1998.

PDF
Distributed optimization and autonomy
We live in an automated world. An increasing number of our daily decisions are guided by computers that sense, infer and act on the data that they observe: search engines monitor our queries and browsing patterns to personalize how they guide us through the vast information on the Internet; automatic trading systems sift through huge amounts of real-time data to identify and exploit market inefficiencies before they disappear; and modern cars detect dangerous situations and temporarily overtake the control from the driver to bring the vehicle back to safety. Autonomous decision-making is soon everywhere and in everything. We develop technology and methodology for making quicker and better informed autonomous decisions, with a particular focus on optimization algorithms for distributed coordination of large networked systems.

Selected publications

Optimal parameter selection for the alternating method of multipliers (ADMM): quadratic problems

E. Ghadimi, A. Teixeira, I. Shames and M. Johansson

IEEE Transactions on Automatic Control, March 2015

PDF

Distributed coordination of household electricity consumption

M. Juelsgaard, A. Teixeira, M. Johansson, R. Wisniewski and J.D. Bendtsen

IEEE Multi-conference on Systems and Control, 2014.

PDF

The evolution of beliefs over signed social networks

G. Shi, A. Proutiere, M. Johansson, J.S. Baras and K.H. Johansson

Operations Research, 2013.

PDF

On the convergence rate of asynchronous iterations

H.R. Feyzmahdavian and M. Johansson

IEEE Conference on Decision and Control, 2014

PDF

Multi-step gradient methods for networked optimization

E. Ghadimi, I. Shames and M. Johansson

IEEE Transactions on Signal Processing, 2013

PDF

On decentralized negotiation of optimal consensus

B. Johansson, A. Speranzon, M. Johansson and K.H. Johansson

Automatica, 2008.

PDF

A randomized incremental subgradient method for distributed optimization in networked systems

B. Johansson, M. Rabi and M. Johansson

SIAM Journal on Optimization, 2009.

PDF

Subgradient methods and consensus algorithms for solving convex optimization problems

B. Johansson, T. Keviczky, K.H. Johansson and M. Johansson

IEEE Conference on Decision and Control, 2008.

PDF
Real-time and sensor networking
Advances in miniturization, computing and wireless communications has enabled the design of autonomous sensing systems that can be embedded into the environment. These wireless sensor networks, formed by tiny battery-operated wireless nodes, find applications in smart buildings, logistics, and environmental monitoring. Our research has focused on improving communication and information processing abilities of such networks, accounting for their limited energy, computate resources and memory. A particular aspect of interacting with the physical world is that information looses value with time: a GPS reading of a car's position taken a minute ago says much less about the current position than a measurement taken a second ago. Similarly, every milisecond of delay from a sensor to a controller in an industrial robot reduces the speed and accuracy at which it can operate. One particular aspect of our research has therefore been devoted to understanding how to transmit data with hard deadlines over unreliable media.

Selected publications

Opportunistic routing in low duty-cycled wireless sensor networks

E. Ghadimi, O. Landsiedel, P. Soldati, S. Duquennoy and M. Johansson

ACM Transactions on Sensor Networks, 2013

PDF

Low power, low delay: opportunistic routing meets duty cycling

O. Landsiedel, E. Ghadimi, S. Duquennoy and M. Johansson

ACM IPSN, 2012

PDF

Deadline-constrained transmission scheduling and data evacuation in WirelessHART networks

P. Soldati, H. Zhang and M. Johansson

European Control Conference, 2009

PDF

Energy-efficient deadline-constrained maximum reliability forwarding in lossy networks

Z. Zou, P. Soldati, H. Zhang and M. Johansson

IEEE Transactions on Wireless Communications, 2012.

PDF

MobiSense: power-efficient micro-mobility in wireless sensor networks

A. Gonga, O. Landsiedel and M. Johansson

IEEE DCOSS, 2011

PDF
Networks and communications
As the demands on our communication infrastructures continue to rise, it becomes increasingly important that they operate with maximum efficiency. In wireless networks, it becomes critical to manage radio resources, such as transmit powers and shared spectrum. In data networks, it becomes essential to understand how the data traffic flows and fluctuates to direct the data from source to destination in the most efficient way. We have worked with both wired and wireless data networks, with a particular focus on the design of distributed resource management mechanisms which allow these networks to operate close to their practical capacity.

Selected publications

Contractive interference functions and rates of convergence of distributed power control laws

H.R. Feyzmahdavian, M. Johansson and T. Charalambous

IEEE Transactions on Wireless Communications, 2012

PDF

Simultaneous routing and resource allocation via dual decomposition

L. Xiao, M. Johansson and S.P. Boyd

IEEE Transactions on Communications, 2004

PDF

Proportionally fair allocation of end-to-end bandwidth in STDMA networks

P. Soldati, B. Johansson and M. Johansson

ACM MobiHoc, 2006

PDF

Mathematical decomposition techniques for distributed cross-layer optimization of data networks

B. Johansson, P. Soldati and M. Johansson

IEEE Journal on Selected Areas in Communications, 2006

PDF

Cross-layer optimization of wireless networks using nonlinear column generation

M. Johansson and L. Xiao

IEEE Transactions on Wireless Communications, 2006

PDF

Traffic matrix estimation on a large IP backbone: a comparison on real data

A. Gunnar, M. Johansson and T. Telkamp

ACM Conference on Internet measurement, 2004

PDF

A survey on buffer-aided relay selection

N. Nomikos. T. Charalambous, I. Krikidis, D.N. Skoutas, D. Vouyioukas, M. Johansson and C. Skianis

IEEE Communications Surveys and Tutorials, 2016

PDF

Energy efficient D2D communications in dynamic TDD systems

D. Della Penda, L. Fu and M. Johansson

IEEE Transactions on Wireless Communications, 2016

PDF

Optimal radio frequency energy harvesting with limited energy arrival knowledge

Z. Zou, A. Gidmark, T. Charalambous and M. Johansson

IEEE Journal on Selected Areas in Communicatoins, 2016

PDF