Penalty-based Method Research Articles

An Improved Penalty-Based Excited-State Variational Monte Carlo Approach with Deep-Learning Ansatzes.

We introduce several improvements to the penalty-based variational quantum Monte Carlo (VMC) algorithm for computing electronic excited states of Entwistle et al. [Nat. Commun. 14, 274 (2023)] and demonstrate that the accuracy of the updated method is competitive with other available excited-state VMC approaches. A theoretical comparison of the computational aspects of these algorithms is presented, where several benefits of the penalty-based method are identified. Our main contributions include an automatic mechanism for tuning the scale of the penalty terms, an updated form of the overlap penalty with proven convergence properties, and a new term that penalizes the spin of the wave function, enabling the selective computation of states with a given spin. With these improvements, along with the use of the latest self-attention-based ansatz, the penalty-based method achieves a mean absolute error below 1 kcal/mol for the vertical excitation energies of a set of 26 atoms and molecules, without relying on variance matching schemes. Considering excited states along the dissociation of the carbon dimer, the accuracy of the penalty-based method is on par with that of natural-excited-state (NES) VMC, while also providing results for additional sections of the potential energy surface, which were inaccessible with the NES method. Additionally, the accuracy of the penalty-based method is improved for a conical intersection of ethylene, with the predicted angle of the intersection agreeing well with both NES-VMC and multireference configuration interaction.

Automated shape and thickness optimization for non-matching isogeometric shells using free-form deformation

AbstractIsogeometric analysis (IGA) has emerged as a promising approach in the field of structural optimization, benefiting from the seamless integration between the computer-aided design (CAD) geometry and the analysis model by employing non-uniform rational B-splines (NURBS) as basis functions. However, structural optimization for real-world CAD geometries consisting of multiple non-matching NURBS patches remains a challenging task. In this work, we propose a unified formulation for shape and thickness optimization of separately parametrized shell structures by adopting the free-form deformation (FFD) technique, so that continuity with respect to design variables is preserved at patch intersections during optimization. Shell patches are modeled with isogeometric Kirchhoff–Love theory and coupled using a penalty-based method in the analysis. We use Lagrange extraction to link the control points associated with the B-spline FFD block and shell patches, and we perform IGA using the same extraction matrices by taking advantage of existing finite element assembly procedures in the FEniCS partial differential equation (PDE) solution library. Moreover, we enable automated analytical derivative computation by leveraging advanced code generation in FEniCS, thereby facilitating efficient gradient-based optimization algorithms. The framework is validated using a collection of benchmark problems, demonstrating its applications to shape and thickness optimization of aircraft wings with complex shell layouts.

A penalty-based method for solving a discrete HJB complementarity problem

A penalty-based method for solving a discrete HJB complementarity problem

QoS-Guaranteed Hybrid Beamforming Design for Multi-User Systems With Finite-Resolution Phase Shifters

Massive multiple input multiple output (MIMO) technology is a key enabler of the next generation wireless communications. The partially-connected hybrid beamforming architecture is attractive in massive MIMO systems for its reduced cost, power consumption, and hardware complexity. On the other hand, the specific analog beamforming structure and finite-resolution in phase shifters complicate the beamforming design. In this paper, we study the hybrid beamforming design for the partially-connected architecture, where users can have multiple antennas and different requirements on quality of service (QoS). A transmit power minimization problem is formulated. The beamforming design needs to be aligned to discrete phase values due to the practical limitations. Such problem is non-convex and is hard to solve. We propose an effective algorithm for this problem. In particular, the penalty-based method is used to decouple the variables. The original problem is decomposed into several sub-problems. The block prox-linear method is applied to solve the sub-problems alternatively with convergence guarantee. We show the effectiveness and robustness of our proposed algorithm via simulations. Our proposed algorithm with low resolution phase shifters achieves better performance with lower transmit power than existing methods.

Energy-Efficient Joint Collaborative and Passive Beamforming for Intelligent-Reflecting-Surface-Assisted Wireless Sensor Networks

Intelligent reflecting surface (IRS) provides a promising technology that can improve the energy efficiency of wireless communications, by building the controllable wireless propagation environment with the utilization of massive passive reflecting elements. Motivated by this, we apply the IRS to assist the data collection based on collaborative beamforming in Wireless Sensor Networks (WSNs). With the objective to maximize the network lifetime, we present an energy efficient joint collaborative and passive beamforming design for an IRS-assisted WSN, which jointly optimizes the beamforming vector for all nodes and phase shifts for IRS’s reflecting elements. To efficiently resolve the joint optimization problem, we develop a penalty dual decomposition (PDD) based algorithm combined with augmented Lagrangian method, penalty-based method and block coordinate descent algorithm, which decomposes the joint optimization problem into four simplified subproblems. Especially in solving IRS’s phase shifts optimization subproblem, we formulate this subproblem as a quadratically constrained quadratic programming (QCQP) problem and further propose a low computational complexity approximate iteration algorithm to resolve this QCQP problem, which iteratively linearizes this QCQP problem to a sequence of approximate programming with linear equality constraints, whose closed-form solution is presented. Theoretical analysis indicates the proposed algorithm can converge to a KKT solution. Simulation results indicate that the IRS-assisted design and proposed algorithm improve the energy efficiency and prolong the network lifetime as compared to alternative algorithms.

Integration of NOMA With Reflecting Intelligent Surfaces: A Multi-Cell Optimization With SIC Decoding Errors

Reflecting intelligent surfaces (RIS) has gained significant attention due to its high energy and spectral efficiency in next-generation wireless networks. By using low-cost passive reflecting elements, RIS can smartly reconfigure the signal propagation to extend the wireless communication coverage. On the other hand, non-orthogonal multiple access (NOMA) has been proven as a key air interface technique for supporting massive connections over limited resources. Utilizing the superposition coding and successive interference cancellation (SIC) techniques, NOMA can multiplex multiple users over the same spectrum and time resources by allocating different power levels. This paper proposes a new optimization scheme in a multi-cell RIS-NOMA network to enhance the spectral efficiency under SIC decoding errors. In particular, the power budget of the base station and the transmit power of NOMA users while the passive beamforming of RIS is simultaneously optimized in each cell. Due to objective function and quality of service constraints, the joint problem is formulated as non-convex, which is very complex and challenging to obtain the optimal global solution. To reduce the complexity and make the problem tractable, we first decouple the original problem into two sub-problems for power allocation and passive beamforming. Then, the efficient solution of each sub-problem is obtained in two-steps. In the first-step of For power allocation sub-problem, we transform it to a convex problem by the inner approximation method and then solve it through a standard convex optimization solver in the second-step. Accordingly, in the first-step of passive beamforming, it is transformed into a standard semi-definite programming problem by successive convex approximation and different of convex programming methods. Then, penalty based method is used to achieve a Rank-1 solution for passive beamforming in second-step. Numerical results demonstrate the benefits of the proposed optimization scheme in the multi-cell RIS-NOMA network.

Gain Without Pain: Recycling Reflected Energy From Wireless-Powered RIS-Aided Communications

In this paper, we investigate and analyze energy recycling for a reconfigurable intelligent surface (RIS)-aided wireless-powered communication network. As opposed to the existing works where the energy harvested by Internet of things (IoT) devices only comes from the power station, IoT devices are also allowed to recycle energy from other IoT devices. In particular, we propose group switching-and user switching-based protocols with time-division multiple access to evaluate the impact of energy recycling on the system performance. Two different optimization problems are respectively formulated for maximizing the sum throughput by jointly optimizing the energy beamforming vectors, the transmit power, the transmission time, the receive beamforming vectors, the grouping factors, and the phase-shift matrices, where the constraints of the minimum throughput, the harvested energy, the maximum transmit power, the phase shift, the grouping, and the time allocation are taken into account. In light of the intractability of the above problems, we respectively develop two alternating optimization-based iterative algorithms by combining the successive convex approximation method and the penalty-based method to obtain corresponding sub-optimal solutions. Simulation results verify that the energy recycling-based mechanism can assist in enhancing the performance of IoT devices in terms of energy harvesting and information transmission. Besides, we also verify that the group switching-based algorithm can obtain more sum throughput of IoT devices, and the user switching-based algorithm can harvest more energy.

Deterministic Optimization Approach for High-Performance Externally Fired Biomass-Fueled Micro-Gas Turbines

Abstract This work focuses on a dual-objective optimization of a 100 kWe externally fired micro-gas turbine utilizing the producer gas from a biomass gasifier. Although externally fired micro-gas turbines are convenient for resolving operability issues in biomass combined heat and power applications, these configurations are still lacking in efficiency compared to the commercial natural-gas fired microturbines. The main cause is the material temperature limitations in the recuperator and the current uneconomical use of high-temperature resistance materials. Toward the achievement of higher efficiency by keeping system economic viability, an optimization process is followed based on the Normal Constraint Method, which generates evenly distributed solutions of a Pareto front. The selected method can determine high-performance solutions, being unidentified by one-dimensional approaches, providing information about the distribution of critical cycle parameters, across the complete objective space by the evaluation of a relatively small set of Pareto points. These critical parameters are the pressure ratio, the recuperator temperature difference, and maximum temperature. The exergetic efficiency and the relative recuperator cost are the optimization objectives. The deterministic Nelder–Mead algorithm is used for the acquisition of Pareto solutions, along with a penalty-based method to perform the constrained optimization. The implemented optimization method can identify superior solutions compared to one-dimensional approaches, as the latter result in higher recuperator costs around 41–112% at the same exergetic efficiency, revealing that high-performance is not only restricted by the recuperator but also by the compressor operating range.

Energy-Efficient Beamforming and Resource Optimization for AmBSC-Assisted Cooperative NOMA IoT Networks

In this manuscript, we present an energy-efficient alternating optimization framework based on the multi-antenna ambient backscatter communication (AmBSC) assisted cooperative non-orthogonal multiple access (NOMA) for next-generation (NG) internet-of-things (IoT) enabled communication networks. Specifically, the energy-efficiency maximization is achieved for the considered AmBSC-enabled multi-cluster cooperative IoT NOMA system by optimizing the active-beamforming vector and power-allocation coefficients (PAC) of IoT NOMA users at the transmitter, as well as passive-beamforming vector at the multi-antenna assisted backscatter node. Usually, increasing the number of IoT NOMA users in each cluster results in inter-cluster interference (ICI) (among different clusters) and intra-cluster interference (among IoT NOMA users). To combat the impact of ICI, we exploit a zero-forcing (ZF) based active-beamforming, as well as an efficient clustering technique at the source node. Further, the effect of intra-cluster interference is mitigated by exploiting an efficient power-allocation policy that determines the PAC of IoT NOMA users under the quality-of-service (QoS), cooperation, SIC decoding, and power-budget constraints. Moreover, the considered non-convex passive-beamforming problem is transformed into a standard semi-definite programming (SDP) problem by exploiting the successive-convex approximation (SCA) approximation, as well as the difference of convex (DC) programming, where Rank-1 solution of passive-beamforming is obtained based on the penalty-based method. Furthermore, the numerical analysis of simulation results demonstrates that the proposed energy-efficiency maximization algorithm exhibits an efficient performance by achieving convergence within only a few iterations.

STAR-RIS-Aided Mobile Edge Computing: Computation Rate Maximization With Binary Amplitude Coefficients

In this paper, simultaneously transmitting and reflecting (STAR) reconfigurable intelligent surface (RIS) is investigated in the multi-user mobile edge computing (MEC) system to improve the computation rate. Compared with traditional RIS-aided MEC, STAR-RIS extends the service coverage from half-space to full-space and provides new flexibility for improving the computation rate for end users. However, the STAR-RIS-aided MEC system design is a challenging problem due to the non-smooth and non-convex binary amplitude coefficients with coupled phase shifters. To fill this gap, this paper formulates a computation rate maximization problem via the joint design of the STAR-RIS phase shifts, reflection and transmission amplitude coefficients, the receive beamforming vectors, and energy partition strategies for local computing and offloading. To tackle the discontinuity caused by binary variables, we propose an efficient smoothing-based method to decrease convergence error, in contrast to the conventional penalty-based method, which brings many undesired stationary points and local optima. Furthermore, a fast iterative algorithm is proposed to obtain a stationary point for the joint optimization problem, with each subproblem solved by a low-complexity algorithm, making the proposed design scalable to a massive number of users and STAR-RIS elements. Simulation results validate the strength of the proposed smoothing-based method and show that the proposed fast iterative algorithm achieves a higher computation rate than the conventional method while saving the computation time by at least an order of magnitude. Moreover, the resultant STAR-RIS-aided MEC system significantly improves the computation rate compared to other baseline schemes with conventional reflect-only/transmit-only RIS.

Enhancing NOMA Networks via Reconfigurable Multi-Functional Surface

By flexibly manipulating the radio propagation environment, reconfigurable intelligent surface (RIS) is a promising technique for future wireless communications. However, the single-side coverage and double-fading attenuation faced by conventional RISs largely restrict their applications. To address this issue, we propose a novel concept of multi-functional RIS (MF-RIS), which provides reflection, transmission, and amplification simultaneously for the incident signal. With the aim of enhancing the performance of a non-orthogonal multiple-access (NOMA) downlink multiuser network, we deploy an MF-RIS to maximize the sum rate by jointly optimizing the active beamforming and MF-RIS coefficients. Then, an alternating optimization algorithm is proposed to solve the formulated non-convex problem by exploiting successive convex approximation and penalty-based method. Numerical results show that the proposed MF-RIS outperforms conventional RISs under different settings.

STAR-RIS Enabled Downlink Secure NOMA Network Under Imperfect CSI of Eavesdroppers

Reconfigurable intelligence surface (RIS) is regarded as one of the key technologies for the future wireless communications. Recently, simultaneous transmission and reflection (STAR) RIS is proposed to overcome the geographical constraint of the conventional reflection only RIS. In this letter, we model the STAR-RIS enabled downlink secure non-orthogonal multiple access systems under imperfect channel state information (CSI) of eavesdroppers. Specifically, we propose the block coordinate descent based optimization framework to maximize the worst secrecy rate with active beamforming of downlink NOMA at the base station and STAR-RIS coefficients, where the uncertainty of eavesdropper’s CSI is transformed by S-Procedure. The successive convex approximation and a penalty-based method are utilized to address the non-convexity of the problem. Finally, simulation results demonstrate that the worst secrecy rate of the proposed algorithm outperforms that of the conventional RIS and the other benchmark schemes.

Accelerating Federated Edge Learning via Topology Optimization

Federated edge learning (FEEL) is envisioned as a promising paradigm to achieve privacy-preserving distributed learning. However, it consumes excessive learning time due to the existence of straggler devices. In this article, a novel topology-optimized FEEL (TOFEL) scheme is proposed to tackle the heterogeneity issue in federated learning and to improve the communication-and-computation efficiency. Specifically, a problem of jointly optimizing the aggregation topology and computing speed is formulated to minimize the weighted summation of energy consumption and latency. To solve the mixed-integer nonlinear problem, we propose a novel solution method of penalty-based successive convex approximation (SCA), which converges to a stationary point of the primal problem under mild conditions. To facilitate real-time decision making, an imitation-learning-based method is developed, where deep neural networks (DNNs) are trained offline to mimic the penalty-based method, and the trained imitation DNNs are deployed at the edge devices for online inference. Thereby, an efficient imitation-learning-based approach is seamlessly integrated into the TOFEL framework. Simulation results demonstrate that the proposed TOFEL scheme accelerates the federated learning process and achieves a higher energy efficiency. Moreover, we apply the scheme to 3-D object detection with multivehicle point cloud data sets in the CARLA simulator. The results confirm the superior learning performance of the TOFEL scheme over conventional designs with the same resource and deadline constraints.

DAE-PINN: a physics-informed neural network model for simulating differential algebraic equations with application to power networks

Deep learning-based surrogate modeling is becoming a promising approach for learning and simulating dynamical systems. However, deep-learning methods find it very challenging to learn stiff dynamics. In this paper, we develop DAE-PINN, the first effective physics-informed deep-learning framework for learning and simulating the solution trajectories of nonlinear differential-algebraic equations (DAE). DAEs are used to model complex engineering systems, e.g., power networks, and present a “form” of infinite stiffness, which makes learning their solution trajectories challenging. Our DAE-PINN bases its effectiveness on the synergy between implicit Runge–Kutta time-stepping schemes (designed specifically for solving DAEs) and physics-informed neural networks (PINN) (deep neural networks that we train to satisfy the dynamics of the underlying problem). Furthermore, our framework (i) enforces the neural network to satisfy the DAEs as (approximate) hard constraints using a penalty-based method and (ii) enables simulating DAEs for long-time horizons. We showcase the effectiveness and accuracy of DAE-PINN by learning the solution trajectories of a three-bus power network.

Intelligent-Reflecting-Surface-Aided Mobile Edge Computing With Binary Offloading: Energy Minimization for IoT Devices

Mobile edge computing (MEC) is envisioned as a promising technique to support computation-intensive and time-critical applications in future Internet of Things (IoT) era. However, the uplink transmission performance will be highly impacted by the hostile wireless channel, the low bandwidth, and the low transmission power of IoT devices. Recently, intelligent reflecting surface (IRS) has drawn much attention because of its capability to control the wireless environments so as to enhance the spectrum and energy efficiencies of wireless communications. In this article, we consider an IRS-aided multidevice MEC system where each IoT device follows the binary offloading policy, i.e., a task has to be computed as a whole either locally or remotely at the edge server. We aim to minimize the total energy consumption of devices by jointly optimizing the binary offloading modes, the CPU frequencies, the offloading powers, the offloading times, and the IRS phase shifts for all devices. Two algorithms, which are greedy based and penalty based, are proposed to solve the challenging nonconvex and discontinuous problem. It is found that the penalty-based method has only linear complexity with respect to the number of devices, but it performs close to the greedy-based method with cubic complexity with respect to the number of devices. Furthermore, binary offloading via IRS indeed saves more energy compared to the case without IRS.

Time-Varying Optimization-Based Approach for Distributed Formation of Uncertain Euler-Lagrange Systems.

We investigate a distributed time-varying formation control problem for an uncertain Euler-Lagrange system. A time-varying optimization-based approach is proposed. Based on this approach, the robots can achieve the expected formation configuration and meanwhile optimize a global objective function using only neighboring and local information. We consider the time-varying optimization where the objective functions can change in real time. In this case, the consensus-based formation tracking control issues and formation containment tracking control issues in the literature can be solved by the proposed approach. By a penalty-based method, the robots' states asymptotically converge to the estimated optimal solution to an equivalent time-varying optimization problem, whose optimal solution can achieve simultaneous formation and optimization. Furthermore, we consider two more general scenarios: 1) the local objective functions can have non-neighbor's information and 2) the optimization problems can have inequality constraints.

A structural reanalysis assisted harmony search for the optimal design of structures

Structural optimization usually requires a large amount of time-consuming structural analyses to evaluate the feasibility of the designs, which obstructs its application in large-scale problems. This study proposed a structural reanalysis-assisted harmony search (RAHS) algorithm for structural optimization. The RAHS utilizes the combined approximation (CA) technique instead of direct finite element analysis to evaluate the performances of the modified structures generated in iterations. Different from existing approaches, the proposed RAHS will update the reanalysis design point set and select the best reanalysis design point based on the similarity of the design vectors to improve the accuracy of the CA in a large search space. The RAHS also develops a diversity-based error criterion to accommodate the accuracy requirements of CA in different search phases of the harmony search. Five benchmark examples are utilized to examine the performances of the proposed RAHS. The results demonstrate that the Cosine distance outperforms the other four design point selection approaches, and the diversity-based error criterion can significantly reduce the computational cost. Compared with the penalty-based method and the improved Deb rule, the number of FE-based structural analyses can be lessened by about 90% and 50%, respectively, while yielding similar optimal designs.

Resource allocation for multi‐IRS‐aided D2D communication underlying cellular networks

This paper investigates the resource allocation design in device-to-device (D2D) communication underlying cellular networks, which is assisted by multiple intelligent reflecting surfaces (IRSs) deployed at the cell boundary to enhance desired signals and mitigate interference between D2D pairs and CUs. In this regard, a multi-objective optimization problem (MOOP) framework is formulated to jointly maximize the downlink sum-rate of the D2D pairs and cellular users (CUs). To doing so, the MOOP is first converted into a single-objective optimization problem (SOOP) by invoking the weighted sum method. Next, to facilitate the high-coupled non-convex SOOP formulation, it is decomposed into two sub-problems through the alternative optimization method. For the first sub-problem, the inner approximation and successive convex approximations are adopted for jointly allocating D2D transmission power, subcarrier assignment, and transmit beamforming matrix at the BS. For the second sub-problem, an iterative algorithm based on the penalty-based method as well as Fenchel's duality approach is adopted to design the passive beamforming matrices at IRSs, which is guaranteed to rapidly converge to a stationary point. Simulation results reveal that a non-trivial trade-off between the total throughput of CUs and D2D pairs exist. Furthermore, the proposed framework significantly outperforms existing works addressed in the literature.

Shortest Path Algorithms for Pedestrian Navigation Systems

Efficient shortest path algorithms are of key importance for routing and navigation systems. However, these applications are designed focusing on the requirements of motor vehicles, and therefore, finding paths in pedestrian sections of urban areas is not sufficiently supported. In addition, finding the shortest path is often not adequate for urban sidewalk routes, as users of these applications may also be interested in alternative routes that, although slightly longer, possess other desirable features and properties. According to the literature, the search for alternative routes is carried out mainly using the k-shortest paths (KSP) algorithm which represents an ordered list of all available alternatives. Even though various KSP algorithms have been proposed, to the best of our knowledge, there is no research addressing all issues inherent in a pedestrian navigation system. The purpose of this paper is to present a heuristic algorithm for graph datasets that implements a penalty-based method which, by increasing certain edge weights, effectively searches for the most accessible alternative paths in multi-route cases. To demonstrate how the algorithm works, we present experimental results on finding the most accessible paths in pedestrian sections of the historical center of Thessaloniki city.

The Framework of Cross-Domain and Model Adversarial Attack against Deepfake

To protect images from the tampering of deepfake, adversarial examples can be made to replace the original images by distorting the output of the deepfake model and disrupting its work. Current studies lack generalizability in that they simply focus on the adversarial examples generated by a model in a domain. To improve the generalization of adversarial examples and produce better attack effects on each domain of multiple deepfake models, this paper proposes a framework of Cross-Domain and Model Adversarial Attack (CDMAA). Firstly, CDMAA uniformly weights the loss function of each domain and calculates the cross-domain gradient. Then, inspired by the multiple gradient descent algorithm (MGDA), CDMAA integrates the cross-domain gradients of each model to obtain the cross-domain perturbation vector, which is used to optimize the adversarial example. Finally, we propose a penalty-based gradient regularization method to pre-process the cross-domain gradients to improve the success rate of attacks. CDMAA experiments on four mainstream deepfake models showed that the adversarial examples generated from CDMAA have the generalizability of attacking multiple models and multiple domains simultaneously. Ablation experiments were conducted to compare the CDMAA components with the methods used in existing studies and verify the superiority of CDMAA.