Luca Schenato

This article presents a second-order fully distributed optimization algorithm, HBNET-GIANT, driven by heavy-ball momentum, for $L$-smooth and $μ$-strongly convex objective functions. A rigorous convergence analysis is performed, and we demonstrate global linear convergence under certain sufficient conditions. Through extensive numerical experiments, we show that HBNET-GIANT with heavy-ball momentum achieves acceleration, and the corresponding rate of convergence is strictly faster than its non-accelerated version, NETWORK-GIANT. Moreover, we compare HBNET-GIANT with several state-of-the-art algorithms, both momentum-based and without momentum, and report significant performance improvement in convergence to the optimum. We believe that this work lays the groundwork for a broader class of second-order Newton-type algorithms with momentum and motivates further investigation into open problems, including an analytical proof of local acceleration in the fully distributed setting for convex optimization problems.

We study optimal proportional feedback controllers for spatially invariant systems when the controller has access to delayed state measurements received from different spatial locations. We analyze how delays affect the spatial locality of the optimal feedback gain leveraging the problem decoupling in the spatial frequency domain. For the cases of expensive control and small delay, we provide exact expressions of the optimal controllers in the limit for infinite control weight and vanishing delay, respectively. In the expensive control regime, the optimal feedback control law decomposes into a delay-aware filtering of the delayed state and the optimal controller in the delay-free setting. Under small delays, the optimal controller is a perturbation of the delay-free one which depends linearly on the delay. We illustrate our analytical findings with a reaction-diffusion process over the real line and a multi-agent system coupled through circulant matrices, showing that delays reduce the effectiveness of optimal feedback control and may require each subsystem within a distributed implementation to communicate with farther-away locations.

We study optimal proportional feedback controllers for spatially invariant systems when the controller has access to delayed state measurements received from different spatial locations. We analyze how delays affect the spatial locality of the optimal feedback gain leveraging the problem decoupling in the spatial frequency domain. For the cases of expensive control and small delay, we provide exact expressions of the optimal controllers in the limit for infinite control weight and vanishing delay, respectively. In the expensive control regime, the optimal feedback control law decomposes into a delay-aware filtering of the delayed state and the optimal controller in the delay-free setting. Under small delays, the optimal controller is a perturbation of the delay-free one, which depends linearly on the delay. We illustrate our analytical findings with a reaction–diffusion process over the real line and a multiagent system coupled through circulant matrices, showing that delays reduce the effectiveness of optimal feedback control and may require each subsystem within a distributed implementation to communicate with farther-away locations.

In wireless networked control systems, a common challenge arises in deciding between a remote controller, with higher computational power but delayed information due to wireless communication, and a local onboard controller, with limited computing resources but timely feedback. This letter investigates such a trade-off in the context of the Linear Quadratic Gaussian (LQG) control, focusing on the interplay between computational power and communication delay. We compare two typical controllers: a local controller operating at a coarse control rate and a remote controller utilizing a finer control rate but experiencing network delay. A formal analytical framework is proposed to quantify the impact of network delay on the LQG cost and computation resources. We show that there exists a critical communication delay above which the more powerful remote controller is no more convenient with respect to the simpler local controller, providing a guideline for controller selection in varying network conditions with fixed computing resource configurations. Our approach provides a lightweight, interpretable selection rule based on the communication-computation trade-off. Simulation results validate the effectiveness of our theoretical insights.

In recent years, federated learning has been widely studied to speed up various supervised learning tasks at the wireless network edge. However, there is a lack of theoretical understanding as to whether similar speedups in sample complexity can be achieved for cooperative reinforcement learning (RL) problems subject to communication constraints. To that end, we study a federated policy evaluation problem over wireless fading channels where, to update model parameters, a central server aggregates local temporal difference (TD) update directions from N agents via analog over-the-air computation (OAC). We refer to this scheme as OAC-FedTD and provide a rigorous finite-time convergence analysis of its performance. Our analysis reveals the impact of the noisy fading channels on the convergence rate and establishes a linear convergence speedup w.r.t. the number of agents. Notably, this is the first non-asymptotic analysis of a cooperative RL setting under wireless channels that jointly considers linear value function approximation, Markovian sampling, and the OAC channel-induced distortions and noise. Our work develops the theoretical foundations that are key for relevant advancements in the analysis and design of federated reinforcement learning algorithms over wireless networks.

Quorum Sensing (QS) is a bacterial cell-to-cell communication mechanism allowing to share information about cell density, to adjust gene expression accordingly. Pathogens leverage QS to coordinate virulence and antimicrobial resistance, leading to distinctive population-level behaviors. To support rational design of synthetic biology strategies counteracting these mechanisms, we first mathematically model and compare two common QS architectures: one based on a single positive feedback loop to auto-induce signal molecule synthesis, the other including an additional positive feedback to increase signal molecule receptors production. Our comprehensive analysis of these QS structures and their equilibria highlights the differences in their bistable and hysteretic behaviors. An extensive sensitivity analysis is then performed, highlighting how parameter variations may lead to phenotype alterations in system behavior. Finally, building on our sensitivity analysis, we mathematically model four distinct QS inhibition strategies -signal molecule degradation, pharmaceutical inhibition, CRISPRi, and RNAi -which lead to the design of Quorum-Quenching (QQ) therapeutic approaches. Despite the underlying complex mechanisms, we demonstrate that the effect of the proposed QQ strategies can be captured by varying specific parameters within the QS models. We numerically analyze how these strategies affect the steady-state behavior of both QS models, identifying critical parameter thresholds for effective QS suppression.

Assisted and autonomous driving are rapidly gaining momentum and will soon become a reality. Artificial intelligence and machine learning are regarded as key enablers thanks to the massive amount of data that smart vehicles will collect from onboard sensors. Federated learning is one of the most promising techniques for training global machine learning models while preserving data privacy of vehicles and optimizing communications resource usage. In this article, we propose vehicular radio environment map federated learning (VREM-FL), a computation-scheduling co-design for vehicular federated learning that combines mobility of vehicles with 5G radio environment maps. VREM-FL jointly optimizes learning performance of the global model and wisely allocates communication and computation resources. This is achieved by orchestrating local computations at the vehicles in conjunction with transmission of their local models in an adaptive and predictive fashion, by exploiting radio channel maps. The proposed algorithm can be tuned to trade training time for radio resource usage. Experimental results demonstrate that VREM-FL outperforms literature benchmarks for both a linear regression model (learning time reduced by 28%) and a deep neural network for semantic image segmentation (doubling the number of model updates within the same time window).

The Energy Management Control Systems present in today’s Building Automation Systems (BASs) typically adopt thermostats, potentially integrated with humidity, CO2, and occupancy sensors, to regulate internal temperature within a predefined range to optimize energy consumption. However direct feedback of occupants’ personal comfort is rarely considered. Differently, this work proposes a new thermal control paradigm, referred to as “humans-in-the-buildings", where individuals can directly signal their discomfort to the energy management system which thus adjusts the HVAC control inputs accordingly. We provide a mathematical formulation for this novel comfort-based control and demonstrate that neither temperature sensors nor knowledge of the occupants’ discomfort profiles is required, as each room’s temperature is regulated based solely on the real-time discomfort signals from individuals. Finally, the effectiveness of the proposed comfort-based control is substantiated through simulations conducted on multiple adjacent rooms established building dynamics modeling.