Fast Optimization Research Articles

Multi-Environment Adaptive Fast Constant False Alarm Detection Algorithm Optimization Strategy

Multi-Environment Adaptive Fast Constant False Alarm Detection Algorithm Optimization Strategy

An unmanned tank combat game driven by FPSO-MADDPG algorithm

With the development of artificial intelligence and unmanned technology, unmanned vehicles have been utilized in a variety of situations which may be hazardous to human beings, even in real battle fields. An intelligent unmanned vehicle can be aware of surrounding situations and make appropriate responding decisions. For this purpose, this paper applies Multi-agent Deep Deterministic Policy Gradient (MADDPG) algorithm for vehicle’s of situation awareness and decision making, inside which a Fast Particle Swarm Optimization (FPSO) algorithm is proposed to calculate the optimal vehicle attitude and position; therefore, an improved deep reinforcement learning algorithm FPSO-MADPPG is formed. A specific advantage function is designed for the FPSO portion, which considers angle, distance, outflanking encirclement. A dedicated reward is designed for the MADPPG portion, which considers key factors like angle, distance, and damage. Finally, FPSO-MADPPG is then used in a combat game to operate unmanned tanks. Simulation results show that our method not only can obtain higher winning rate, but also higher reward and faster convergence than DDPG and MADPPG algorithms.

Opinions on fast distributed optimization for large-scale scheduling of heterogeneous flexibility resources

Opinions on fast distributed optimization for large-scale scheduling of heterogeneous flexibility resources

A digital twin for advancing battery fast charging based on a Bayesian optimization-based method

The optimization of fast charging protocols is regarded as a key technology for promoting the use of electric vehicles because it can balance battery charging time and health. Optimizing such charging protocols through electrochemical models is a mainstream approach and can demonstrate high accuracy in simulating battery characteristics. However, the high complexity of the corresponding models makes the calculation and optimization processes difficult to perform in real time. To address this problem, this study presents an online closed-loop fast charging strategy optimization scheme that combines a battery digital twin model (BDTM) and Bayesian optimization (BO). Because the BO can quickly find a near-optimal solution in a few iterations, the optimization time is shortened, thereby reducing the computational burden incurred by highly complex models. To further improve the efficiency of the online optimization, a parallel multichannel optimization strategy is proposed, which further accelerates the process of finding the optimal protocol by simultaneously executing multiple optimization algorithms. Additionally, we analysed the effects of ageing parameters and ambient temperature on the optimization results. The results show that BO can obtain relatively stable optimization and has the highest efficiency when using four parallel channels. Specifically, the number of convergence evaluations for single-channel optimization is 2.5 times that of the four-channel optimization. Compared with the reference charging protocol, charging using the protocol optimized based on the Doyle-Fuller-Newman (DFN) model can effectively suppress the loss of lithium inventory (LLI) by up to 4.76 % within 90 cycles.

Need For Speed: Fast Wind Farm Optimization

The Wind in my Backyard (WIMBY) project is developing a web interface to aid communities in siting wind energy projects. As part of this siting tool, users will be able to find realistic wind farm layouts for any proposed site in Europe, given certain constraints. When designing this tool, there arises a need for speed: realistic layouts must be designed in computational times that are appropriate for a web interface. In this study, we compare two optimization algorithms: a gradient-based algorithm, referred to as stochastic gradient descent (SGD), and a gradient-free method, referred to as smart-start. The trade-offs between the optimal energy yield and optimization computational time are characterized via a parameter sweep, considering a site in Denmark. This analysis considered farms with 10, 25, and 50 turbines. We find that smart-start yielded the best results for very short computational times, and that SGD yielded layouts with higher energy yields when considering larger computational times.

Time-varying neurodynamic optimization approaches with fixed-time convergence for sparse signal reconstruction

Convergence speed is one of the most considerable features for sparse signal reconstruction algorithms. This paper introduces several novel fast time-varying neurodynamic optimization approaches (TVNOAs) with fixed-time convergence for sparse signal reconstruction. It is demonstrated that these trajectories converge to solutions within a fixed-time from arbitrary initial conditions, exhibiting a faster convergence rate due to the selection of appropriate time-varying coefficients. The fixed-time convergence of the proposed TVNOAs is investigated using the Polyak–Lojasiewicz condition. Moreover, explicit upper bounds for the settling time of TVNOAs are provided. Additionally, the robustness under bounded noises is further examined. Numerical experiments on sparse signal reconstruction validate the superiority of the proposed neurodynamic approaches.

Decoding electroencephalographic responses to visual stimuli compatible with electrical stimulation.

Electrical stimulation of the visual nervous system could improve the quality of life of patients affected by acquired blindness by restoring some visual sensations, but requires careful optimization of stimulation parameters to produce useful perceptions. Neural correlates of elicited perceptions could be used for fast automatic optimization, with electroencephalography as a natural choice as it can be acquired non-invasively. Nonetheless, its low signal-to-noise ratio may hinder discrimination of similar visual patterns, preventing its use in the optimization of electrical stimulation. Our work investigates for the first time the discriminability of the electroencephalographic responses to visual stimuli compatible with electrical stimulation, employing a newly acquired dataset whose stimuli encompass the concurrent variation of several features, while neuroscience research tends to study the neural correlates of single visual features. We then performed above-chance single-trial decoding of multiple features of our newly crafted visual stimuli using relatively simple machine learning algorithms. A decoding scheme employing the information from multiple stimulus presentations was implemented, substantially improving our decoding performance, suggesting that such methods should be used systematically in future applications. The significance of the present work relies in the determination of which visual features can be decoded from electroencephalographic responses to electrical stimulation-compatible stimuli and at which granularity they can be discriminated. Our methods pave the way to using electroencephalographic correlates to optimize electrical stimulation parameters, thus increasing the effectiveness of current visual neuroprostheses.

Future role of wave power in Seychelles: A structured sensitivity analysis empowered by a novel EnergyPLAN-based optimisation tool

Mitigating climate change requires a variety of energy technologies and energy simulation approaches to evaluate the best possible system structures. Screening whether novel technologies are a viable solution for a particular country within a cost-optimised system setup is usually simulation- and time-intensive. This study introduces the novel add-on optimisation tool EP-ALISON-LUT for use in combination with EnergyPLAN applied to the test case of wave power in the case of Seychelles in 2030 and 2050 within a structured sensitivity analysis. The tool enables a high number of possible system setups and scenarios, including the import and domestic production of electricity-based fuels, to be modelled, allowing for an in-depth view of the system impacts of integrating wave power. The results indicate a limited role for wave power due to its relatively low yield, especially in 2030. However, in 2050, up to 500 MW of wave power capacity is possible with a lower or similar levelised cost of final energy compared to the reference scenario in 2019, which can benefit the diversification of the power generation portfolio. Thus, this novel tool is fast and effective in technology screening studies requiring a fast optimisation algorithm.

Additional Acceleration of Antenna Optimal Characterizationwith Modeling Support

Optimal characterization is assumed to provide the best solution for the designed cost function among the possible solutions within the specified range. These processes can take a long time depending on the applications and computer hardware used. Here, the optimization process is supported by ANN modeling in order to shorten the current optimization processes as much as possible. For this purpose, the selection of design parameters of the bowtie patch antenna (BPA) is presented as a multi-dimensional, multi-objective modeling-supported design optimization problem. The operating frequency of the proposed antenna is 28 GHz, which is the standard for millimeter wave band and 5G technologies. To overcome this challenging design optimization, a new, fast and powerful optimization algorithm was used by modifying the non-dominant sorting genetic algorithm (NSGA)-III, and the optimal characterization of the microwave antenna design was achieved. Although the proposed method gives the same results compared to the existing process, it takes much less time. Therefore, it is possible to shorten the process and reduce costs without the need for extra applications or hardware. As a whole, the proposed design optimization process is an efficient, fast and reliable solution for all design problems.

Fast Convex Optimization via Differential Equation with Hessian-Driven Damping and Tikhonov Regularization

Fast Convex Optimization via Differential Equation with Hessian-Driven Damping and Tikhonov Regularization

Alrecon: computed tomography reconstruction web application based on Solara

Synchrotron X-ray computed tomography is a non-destructive 3D imaging technique that offers the possibility to study the internal microstructure of samples with high spatial and temporal resolution. Given its unmatched image quality and acquisition speed, and the possibility to preserve the specimens, there is an increasing demand for this technique, from scientific users from innumerable disciplines. Computed tomography reconstruction is the computational process by which experimental radiographs are converted to a meaningful 3-dimensional image after the scan. The procedure involves pre-processing steps for image background and artifact correction on raw data, a reconstruction step approximating the inverse Radon-transform, and writing of the reconstructed volume image to disk. Several open-source Python packages exist to help scientists in the process of tomography reconstruction, by offering efficient implementations of reconstruction algorithms exploiting central or graphics processing unit (CPU and GPU, respectively), and by automating significant portions of the data processing pipeline. A further increase in productivity is attained by scheduling and parallelizing demanding reconstructions on high performance computing (HPC) clusters. Nevertheless, visual inspection and interactive selection of optimal reconstruction parameters remain crucial steps that are often performed in close interaction with the end-user of the data. As a result, the reconstruction task involves more than one software. Graphical user interfaces are provided to the user for fast inspection and optimization of reconstructions, while HPC resources are often accessed through scripts and command line interface. We propose Alrecon, a pure Python web application for tomographic reconstruction built using Solara. Alrecon offers users an intuitive and reactive environment for exploring data and customizing reconstruction pipelines. By leveraging upon popular 3D image visualization tools, and by providing a user-friendly interface for reconstruction scheduling on HPC resources, Alrecon guarantees productivity and efficient use of resources for any type of beamline user.

Penalty weight tuning in high dose rate brachytherapy using multi-objective Bayesian optimization

Objective. Treatment plan optimization in high dose rate brachytherapy often requires manual fine-tuning of penalty weights for each objective, which can be time-consuming and dependent on the planner's experience. To automate this process, this study used a multi-criteria approach called multi-objective Bayesian optimization with q-noisy expected hypervolume improvement as its acquisition function (MOBO-qNEHVI). Approach. The treatment plans of 13 prostate cancer patients were retrospectively imported to a research treatment planning system, RapidBrachyMTPS, where fast mixed integer optimization (FMIO) performs dwell time optimization given a set of penalty weights to deliver 15 Gy to the target volume. MOBO-qNEHVI was used to find patient-specific Pareto optimal penalty weight vectors that yield clinically acceptable dose volume histogram metrics. The relationship between the number of MOBO-qNEHVI iterations and the number of clinically acceptable plans per patient (acceptance rate) was investigated. The performance time was obtained for various parameter configurations. Main results. MOBO-qNEHVI found clinically acceptable treatment plans for all patients. With increasing the number of MOBO-qNEHVI iterations, the acceptance rate grew logarithmically while the performance time grew exponentially. Fixing the penalty weight of the tumour volume to maximum value, adding the target dose as a parameter, initiating MOBO-qNEHVI with 25 parallel sampling of FMIO, and running 6 MOBO-qNEHVI iterations found solutions that delivered 15 Gy to the hottest 95% of the clinical target volume while respecting the dose constraints to the organs at risk. The average acceptance rate for each patient was 89.74% ± 8.11%, and performance time was 66.6 ± 12.6 s. The initiation took 22.47 ± 7.57 s, and each iteration took 7.35 ± 2.45 s to find one Pareto solution.Significance. MOBO-qNEHVI combined with FMIO can automatically explore the trade-offs between treatment plan objectives in a patient specific manner within a minute. This approach can reduce the dependency of plan quality on planner’s experience and reduce dose to the organs at risk.

Biological dose optimization incorporating intra-tumoural cellular radiosensitivity heterogeneity in ion-beam therapy treatment planning

Objective. Treatment plans of ion-beam therapy have been made under an assumption that all cancer cells within a tumour equally respond to a given radiation dose. However, an intra-tumoural cellular radiosensitivity heterogeneity clearly exists, and it may lead to an overestimation of therapeutic effects of the radiation. The purpose of this study is to develop a biological model that can incorporate the radiosensitivity heterogeneity into biological optimization for ion-beam therapy treatment planning. Approach. The radiosensitivity heterogeneity was modeled as the variability of a cell-line specific parameter in the microdosimetric kinetic model following the gamma distribution. To validate the developed intra-tumoural-radiosensitivity-heterogeneity-incorporated microdosimetric kinetic (HMK) model, a treatment plan with H-ion beams was made for a chordoma case, assuming a radiosensitivity heterogeneous region within the tumour. To investigate the effects of the radiosensitivity heterogeneity on the biological effectiveness of H-, He-, C-, O-, and Ne-ion beams, the relative biological effectiveness (RBE)-weighted dose distributions were planned for a cuboid target with the stated ion beams without considering the heterogeneity. The planned dose distributions were then recalculated by taking the heterogeneity into account. Main results. The cell survival fraction and corresponding RBE-weighted dose were formulated based on the HMK model. The first derivative of the RBE-weighted dose distribution was also derived, which is needed for fast biological optimization. For the patient plan, the biological optimization increased the dose to the radiosensitivity heterogeneous region to compensate for the heterogeneity-induced reduction in biological effectiveness of the H-ion beams. The reduction in biological effectiveness due to the heterogeneity was pronounced for low linear energy transfer (LET) beams but moderate for high-LET beams. The RBE-weighted dose in the cuboid target decreased by 7.6% for the H-ion beam, while it decreased by just 1.4% for the Ne-ion beam. Significance. Optimal treatment plans that consider intra-tumoural cellular radiosensitivity heterogeneity can be devised using the HMK model.

Transfer learning of convolutional neural network model for thermal estimation of multichip modules

This paper proposes a transfer learning approach to reduce the dependence of the neural network model on dataset size for multi-chip modules (MCMs) thermal estimation. The dataset that resembles the target task is used to pre-train a convolutional neural network model; then, it is fine-tuned using a much smaller dataset belonging to the target task. This process enables the model to adapt and refine its thermal estimation capabilities for the particular MCM under consideration with sparse data. Results show that the model effectively utilizes the geometric shapes and power details of MCMs to precisely predict steady-state temperature fields and junction temperatures. In a complex five-chip configuration, the fine-tuned model achieves a prediction error of only 1.17 % and accelerates inference speed by three orders of magnitude compared to traditional simulations. Additionally, the transfer learning model requires only 1/8 the dataset size and less than 1/3 the training time compared to a model with random initial parameters to achieve similar accuracy. The results indicate that this model can significantly enhance the utilization efficiency of historically accumulated MCM heat transfer datasets for new demands such as real-time thermal estimation and fast optimization of chip configuration.

Fast robust optimization of ORC based on an artificial neural network for waste heat recovery

Uncertainty in the organic Rankine cycle (ORC) system and the highly complex ORC models pose challenges to optimal operation. A data-driven robust parametric optimization for the ORC system is proposed to ensure high and stable performance under uncertainty. The uncertainty of the cyclic variables is estimated by their distribution parameters, and the average (expected) thermodynamic performance (the net output power of the ORC) is maximized as the optimization objective while minimizing its variance. An artificial neural network with the rectified linear unit is used as a surrogate model for optimization. Then the robust parametric optimization problem can be transformed into a mixed-integer linear optimization problem with a chance-constrained form, and the robust optimal solution can be solved quickly. Case studies show that the solution time of the robust optimization problem using the proposed method is about 1.4 s, which is much less than obtaining the optimal solution based on ORC mechanism models. Meanwhile, the optimal operating condition derived by the proposed robust optimization approach outperforms that obtained by the traditional deterministic strategy. The proposed approach not only improves the robustness of the system but also demonstrates the importance of considering uncertainty in parametric optimization.

Lissajous curves as aerial search patterns

Manned and unmanned systems are prevalent in a wide range of aerial searching applications. For aircraft whose trajectory is not or cannot be planned on-the-fly, optimal deterministic search pattern generation is a critical area of research. Lissajous curves have recently caught attention as excellent candidates for all kinds of aerial search applications, but little fundamental research has been done to understand how best to design Lissajous pattern (LP)s for this use. This paper examines the optimization of these search patterns from analytical, numerical, and data-driven perspectives to establish the state of the field in Lissajous curves for aerial search. From an analytical perspective, it was found that the average expected distance between a Lissajous searcher and a random target on a unit square approaches 0.586 as search time increases. Furthermore, an analytical approximation for the average searcher speed was found to guarantee error of no more than 22.1%. Important outcomes from the numerical optimization of Lissajous search patterns include the development of an intuitive evaluation criterion and the conclusion that irrational frequency ratios near 0.8 typically yield highest performance. Finally, while a robust predictive model for fast pattern optimization is yet out of reach, initial results indicate that such an approach shows promise.

Fast Prediction and Optimization of Building Wind Environment Using CFD and Deep Learning Method

CFD offers advantages over wind tunnel experiments in the prediction and optimization of building wind environment; however, the computational costs associated with optimizing architectural wind environment remain a challenge. In this study, an approach that combines deep learning techniques with CFD simulations is proposed for the prediction and optimization of the architectural wind environment efficiently. A dataset of wind field is constructed using CFD simulation, considering various wind directions, wind speeds, and building spacing. Subsequently, a U-net deep learning model is trained as a surrogate model to rapidly predict the architectural wind field under different conditions. The results indicate that the model can accurately predict the wind field in buildings. The prediction time of building wind field is only 1/900 of that of CFD simulations, making it a viable surrogate model for wind environment optimization. Furthermore, considering all the building layouts and inflow conditions examined in this study, the maximum and minimum uniform wind speed area ratios Auni are 0.84 and 0.13, respectively. Under a single inflow speed, the maximum improvement in the Auni is 0.4, with an improvement rate of 48%. The results demonstrate the effectiveness of the proposed method as an efficient approach for optimizing architectural wind environment.

Fast Buffet-Onset Prediction and Optimization Method Based on Pretrained Flowfield Prediction Model

A transonic buffet is a detrimental phenomenon that occurs on supercritical airfoils and limits the aircraft operating envelope. Traditional methods for predicting buffet onset rely on multiple computational fluid dynamics simulations to assess a series of airfoil flowfields and then apply criteria to them, which is slow and hinders optimization efforts. This study introduces an innovative approach for rapid buffet-onset prediction. A machine-learning flowfield prediction model was pretrained on a large database and then deployed offline to replace the simulations in the buffet prediction process for new airfoil designs. Unlike using a model to directly predict buffet onset, the proposed technique offers better visualization capabilities by providing users with intuitive flowfield outputs. It also demonstrates superior generalization ability, as evidenced by a 32.5% reduction in the average buffet-onset prediction error on the testing dataset. This method was used to optimize the buffet performance of 11 distinct airfoils within and outside the training dataset. The optimization results were verified with simulations and proved to yield improved samples across all cases. It was affirmed that the pretrained flowfield prediction model can be applied to accelerate aerodynamic shape optimization, but further work is still needed to raise its reliability for this safety-critical task.

Physical Prior Mean Function-Driven Gaussian Processes Search for Minimum-Energy Reaction Paths with a Climbing-Image Nudged Elastic Band: A General Method for Gas-Phase, Interfacial, and Bulk-Phase Reactions.

The climbing-image nudged elastic band (CI-NEB) method serves as an indispensable tool for computational chemists, offering insight into minimum-energy reaction paths (MEPs) by delineating both transition states (TSs) and intermediate nonstationary structures along reaction coordinates. However, executing CI-NEB calculations for reactions with extensive reaction coordinate spans necessitates a large number of images to ensure a reliable convergence of the MEPs and TS structures, presenting a computationally demanding optimization challenge, even with mildly costly electronic-structure methods. In this study, we advocate for the utilization of physically inspired prior mean function-based Gaussian processes (GPs) to expedite MEP exploration and TS optimization via the CI-NEB method. By incorporating reliable prior physical approximations into potential energy surface (PES) modeling, we demonstrate enhanced efficiency in multidimensional CI-NEB optimization with surrogate-based optimizers. Our physically informed GP approach not only outperforms traditional nonsurrogate-based optimizers in optimization efficiency but also on-the-fly learns the reaction path valley during optimization, culminating in significant advancements. The surrogate PES derived from our optimization exhibits high accuracy compared to true PES references, aligning with our emphasis on leveraging reliable physical priors for robust and efficient posterior mean learning in GPs. Through a systematic benchmark study encompassing various reaction pathways, including gas-phase, bulk-phase, and interfacial/surface reactions, our physical GPs consistently demonstrate superior efficiency and reliability. For instance, they outperform the popular fast inertial relaxation engine optimizer by approximately a factor of 10, showcasing their versatility and efficacy in exploring reaction mechanisms and surface reaction PESs.

Performance evaluation of multilevel multiphase inverters by piece-wised linearized simulation models

The paper deals with the design issue of the computer simulation models of the traction inverter reflecting request on the rapid investigation of the efficiency performance under various operational conditions. According to this, a piece-wised linearized simulation tool PLECS is selected for model development, which considers the relevant loss behaviour of the main power circuit semiconductor components. The topologies, for which the simulation models are created, are topologies, one of which is commonly used across the automotive industry as the heart of the traction drive (VSI – Voltage Source Inverter), and its perspective alternative three-level traction inverter topology T-NPC – T-type Neutral Point Clamped. The simulation models created in the given simulation tool are thus subjected to the analysis of power parameters, which also include the calculated power losses of the entire system of semiconductor devices for both topologies of the traction inverter. Based on the obtained power parameters, the efficiency performance under various operational conditions of both topologies is evaluated. The presented methodology enables rapid investigation of operational performance enabling fast optimization methods. Any design of linearized simulation models in the PLECS simulation tool, whose core consists of semiconductor elements, can be analyzed by the proposed methodology.