Robust Multi-Robot Trajectory Optimization Using Alternating Direction Method of Multiplier

The Cox proportional hazards model is a popular semi-parametric model for survival analysis. In this paper, we aim at developing a federated algorithm for the Cox proportional hazards model over vertically partitioned data (i.e., data from the same patient are stored at different institutions). We propose a novel algorithm, namely VERTICOX, to obtain the global model parameters in a distributed fashion based on the Alternating Direction Method of Multipliers (ADMM) framework. The proposed model computes intermediary statistics and exchanges them to calculate the global model without collecting individual patient-level data. We demonstrate that our algorithm achieves equivalent accuracy for the estimation of model parameters and statistics to that of its centralized realization. The proposed algorithm converges linearly under the ADMM framework. Its computational complexity and communication costs are polynomially and linearly associated with the number of subjects, respectively. Experimental results show that VERTICOX can achieve accurate model parameter estimation to support federated survival analysis over vertically distributed data by saving bandwidth and avoiding exchange of information about individual patients. The source code for VERTICOX is available at: https://github.com/daiwenrui/VERTICOX.

Swarm trajectory optimization problems are a well-recognized class of multi-agent optimal control problems with strong nonlinearity. However, the heuristic nature of needing to set the final time for agents beforehand and the time-consuming limitation of the significant number of iterations prohibit the application of existing methods to large-scale swarms of unmanned aerial vehicles (UAVs) in practice. In this paper, we propose a spatial-temporal trajectory optimization framework that accomplishes multi-UAV consensus based on the alternating direction multiplier method (ADMM) and uses differential dynamic programming (DDP) for fast local planning of individual UAVs. The introduced framework is a two-level architecture that employs parameterized DDP as the trajectory optimizer for each UAV and ADMM to satisfy the local constraints and accomplish the spatial-temporal parameter consensus among all UAVs. This results in a fully distributed algorithm called distributed parameterized DDP. In addition, an adaptive tuning criterion based on the spectral gradient method for the penalty parameter is proposed to reduce the number of algorithmic iterations. Several simulation examples are presented to verify the effectiveness of the proposed algorithm.

Distributed robust operational optimization of networked microgrids embedded interconnected energy hubs

Park-level integrated energy internet (PEI), a multiple energy coupling and obvious zoning system, is set up to meet the growing demand of energy consumption and improve the comprehensive utilization efficiency of multiple energy resources. Traditional distributed alternating direction method of multipliers (ADMM) algorithm with a control center holding all information is often used to solve the dispatch model of PEI, which may cause security problems. In this article, we use the parallel ADMM algorithm to solve the problem in a distributed way. Meanwhile, a completely distributed scheme based on the average-consensus parallel ADMM algorithm is designed to eliminate the computing center of the subregion. The simulation results verified the validity of the proposed method.

Complex ADMM-Net, a complex-valued neural network architecture inspired by the alternating direction method of multipliers (ADMM), is designed for interference removal in super-resolution stepped frequency radar angle-range-doppler imaging. Tailored to an uncooperative scenario wherein a MIMO radar shares spectrum with communications, the ADMM-Net recovers the radar image---which is assumed to be sparse---and simultaneously removes the communication interference, which is modeled as sparse in the frequency domain owing to spectrum underutilization. The scenario motivates an $\ell_1$-minimization problem whose ADMM iteration, in turn, undergirds the neural network design, yielding a set of generalized ADMM iterations that have learnable hyperparameters and operations. To train the network we use random data generated according to the radar and communication signal models. In numerical experiments ADMM-Net exhibits markedly lower error and computational cost than ADMM and CVX.

Decentralized transactive energy management of multi-microgrid distribution systems based on ADMM

ABSTRACTThe advancement of technologies in combined cooling, heat, and power (CCHP) systems has significantly driven the development of energy hubs (EHs), offering an efficient platform to integrate diverse energy carriers. This study proposes a comprehensive framework for interconnected energy hubs within networked microgrids (NMGs), aiming to improve the efficiency of multi‐carrier energy systems through optimal coordination among the involved entities. In the proposed model, both microgrids and the distribution network are treated as autonomous agents, each managing its own local scheduling challenges. To achieve seamless coordination between these entities, a distributed approach based on the alternating direction method of multipliers (ADMM) is utilised to mitigate power exchange conflicts between multiple microgrids and the distribution network. Additionally, energy systems often operate under uncertain conditions due to the integration of renewable energy sources (RES) and their interactions with the electricity market. To address these uncertainties, this study introduces a novel decentralised optimisation framework that integrates robust optimization (RO) with information gap decision theory (IGDT). The proposed model is evaluated on multi‐carrier energy systems within networked microgrids, and the results demonstrate its low computational complexity and guaranteed convergence of the distributed algorithm.

The network identification plays a very prominent role for the network operator to accomplish the various objectives such as state-estimation, monitoring, control, planning, and real-time analytics. The network structure varies from time-to-time and its details are often not available with the network operator. To address this issue, an alternating direction method of multipliers (ADMM) based framework is presented herein to identify the network topology and line parameters using smart meter (SM) and micro phasor measurement unit (micro-PMU) measurements. The presented algorithm is divided into two sections, 1) approximate parameter evaluation through regression, to extract the partial topology information, and 2) complete network topology identification through the ADMM framework. This algorithm accomplishes the objectives of identifying the network configuration, branch parameters (e.g. conductance and susceptance), and change in branch parameters. Simulation results demonstrate the effectiveness of the presented algorithm on the benchmarked IEEE 13-bus and IEEE 123-bus feeders under various operating scenarios. Furthermore, the presented framework illustrates excellent network identification even with the presence of the stochastic nature of renewable power generation. The presented algorithm exhibits an excellent performance even with the consideration of noise in both measurements. In addition, the comparative performance is carried out on the benchmarked unbalanced IEEE 13-bus and balanced IEEE 33-bus feeders to highlight the efficacy of the presented framework over the state-of-art framework.

Poisson image deblurring with frame-based nonconvex regularization

A low-complexity ADMM-based secure beamforming for IRS-assisted uplink cognitive radio networks with NOMA

Synchronous and asynchronous parallel computation for large-scale optimal control of connected vehicles

This article develops a fully decentralized approximate dynamic programming (FD-ADP) algorithm for stochastic energy management (SEM) of a networked microgrid (NMG) system. First, considering the ac power flow constraints, an alternating direction method of multipliers (ADMM)-based decentralized SEM framework is proposed for NMG coordination. Then, a transactive energy scheme is introduced to further decouple each microgrid (MG) optimization for privacy enhancement and computation reduction. Next, a FD-ADP algorithm is proposed to cope with the real-time uncertainties. The piecewise linear function (PLF) is employed for value function approximation, and a fully decentralized PLF slope update method based on ADMM framework is designed for decentralized property preservation, which trains the value function just through each MG local information and neighboring communication, thus the well-trained decentralized PLF slopes can help achieve the global optimal SEM strategy for NMG coordination under stochastic environments. Finally, case studies demonstrate the effectiveness of the proposed ADMM-based FD-ADP algorithm in terms of decentralized optimization, decentralized training, and global optimality.

SummaryExtreme learning machine (ELM) is a prominent example of neural network with its fast training speed, and good prediction performance. An online version of ELM called online sequential extreme learning machine (OS‐ELM) has also been proposed for the sequential training. Combined with the need for regularization to prevent over‐fitting in addition to the large number of neurons required in the hidden layer, OS‐ELM demands huge amount of computation power for the large‐scale data. In this article, a mixed norm () regularized online machine learning algorithm (MRO‐ELM) that is based on alternating direction method of multipliers (ADMM) is proposed. A linear combination of the mixed norm and the Frobenius norm regularization is applied using the ADMM framework and update formulas are derived. Graphics processing unit (GPU) accelerated version of MRO‐ELM (GPU‐MRO‐ELM) is also proposed to reduce the training time by processing appropriate parts in parallel using the implemented custom kernels. In addition, a novel automatic hyper‐parameter tuning method is incorporated to GPU‐MRO‐ELM using progressive validation with GPU acceleration. The experimental results show that the MRO‐ELM algorithm and its GPU version outperform OS‐ELM in terms of training speed, and testing accuracy. Also, compared to the cross validation, the proposed automatic hyper‐parameter tuning demonstrates dramatical reduction in the tuning time.

GCN-based proximal unrolling matrix completion for piecewise smooth signal recovery

Grid-interactive efficient buildings can provide many benefits for the power grid. However, it is very challenging to coordinate their operations intelligently due to many considerations, e.g., privacy protection and scalability. To this end, this paper investigates an optimal coordination operation problem of multiple grid-interactive residential buildings under uncertainties to minimize the total peak demand violation of all residential buildings during the grid service periods while minimizing their respective operation cost. Taking into account multi-source uncertainties, temporal and spatial couplings, inexplicit thermal dynamics models, and the lack of information sharing among residential buildings incurred by privacy protection, it is difficult to design a centralized model-based optimization operation method. To overcome the challenge, we design an online distributed coordination operation algorithm based on the alternating direction method of multipliers (ADMM) and relative entropy regularized policy iteration (RERPI). The designed algorithm consists of two stages, i.e., independent offline training of each RERPI-based agent related to the building energy management system controller (BEMSC) in each residential building (stage-1) and online distributed execution for coordination between the aggregator controller and each BEMSC under the ADMM framework (stage-2). Simulation results verify the effectiveness and scalability of the designed algorithm.