Efficient multi-response adaptive sampling algorithm for construction of variable-fidelity aerodynamic tables

This paper proposed an efficient and adaptive frequency sampling algorithm for frequency response analysis using dynamic condensation-based reduced-order modeling. For the degree of freedom-based model reduction method, the reduced-order basis becomes a frequency-dependent matrix since the relationship between master and slave degrees of freedom stems from partial equations of a second-order dynamical system. Such frequency-dependency makes the analysis inefficient for investigating the frequency response of the system. Considering that the coverage of a local reduced-order basis at a single frequency varies depending on the frequency, a new frequency sampling algorithm was proposed with a strategy of constructing multiple local reduced-order models (ROMs) at sample frequencies. For adaptive sampling, the frequency range of a local ROM was evaluated, and a new sample was added if there was a gap between two adjacent ROMs. As a result, the accuracy of the local ROM can be estimated, and the efficiency in the online stage was greatly enhanced. The proposed method was verified by performing frequency response analysis with several numerical examples, including a large-scale structural and dynamic system.

In a world where materials are becoming more and more scarce, it is of key importance to produce with the least amount of waste as possible. This is also the case for metal forming processes. Unfortunately, metal forming processes are subjected to uncertainty, such as variations in the raw material or the production environment. These disturbances must be taken into account when designing the process in order to manufacture products with high accuracy and low scrap rates. The procedure of finding the design parameters that lead to the least variation in the final product despite the influence of the uncertain parameters, is referred to as robust optimization. To reduce the time and costs associated with process development using physical experiments, computational models are often used. <br/><br/>Metal forming processes are commonly modelled with transient Finite Element Analysis (FEA), which is computationally expensive, with model evaluation times usually in the order of minutes or hours. In general, optimization requires many evaluations of the objective function. Optimizing the production process based on the FE-model is therefore not computationally feasible. Surrogate models are commonly used to overcome this computational burden. A surrogate model is a cheap-to-evaluate model that mimics the output of the expensive model and is constructed using a data set of results from a limited number of evaluations of the expensive model.<br/><br/>In this work, methods for surrogate modelling and robust optimization of multi-stage metal forming processes using computational models are studied. The dissertation focuses on the construction of surrogate models that describe the entire output field of FE-models. Special attention is paid on how the available data should be preprocessed to get as much information as possible from them when different phyiscal quantities need to be modelled. Furthermore it is investigated how Radial Basis Function interpolation can be applied to get the best interpolation of preprocessed data. The obtained knowledge is applied in order to construct a surrogate model of a production stage of a two-stage metal forming process. The output from the surrogate model of the first stage is propagated to the FE-model of the subsequent stage and compared with the output of a surrogate model that describes both stages at once. Lastly, an adaptive sampling strategy for scalar metamodel-based robust optimization with multi-stage models is presented. For this purpose an algorithm is proposed that takes account of the simulation time of different stages.<br/>

Sampling is a core process in IoT systems. It deter-mines the data volume circulating within the network as well as the energy consumption on the IoT devices. Adaptive sampling aims to control the volume of generated data to reduce energy and bandwidth consumption without undermining data quality. Within this context, we propose two new adaptive sampling techniques: a light-weight adaptive sampling algorithm and an optimized uniform sampling method. We tested our methods using various real data-sets and compared their performances against state-of-the-art adaptive sampling algorithms in terms of data quality and data volume. The results show that the proposed methods are consistently among the best with a noticeable reduction in computational load.

Frequency domain sweeps of array antennas are well-known to be time-intensive, and different surrogate models have been used to improve the performance. Data-driven model order reduction algorithms, such as the Loewner framework and vector fitting, can be integrated with adaptive frequency sampling algorithms, in an iterative scheme, to reduce the number of full-wave simulations required to accurately capture the requested frequency behavior of multiport array antennas. In this work, we propose two novel adaptive methods exploiting a block matrix function which is a key part of the Loewner framework generating system approach. The first algorithm leverages an inherent matrix parameter freedom in the block matrix function to identify frequency points with large errors, whereas the second utilizes the condition number of the block matrix function. The first method effectively provide a frequency domain error estimator, which is essential for improved performance. Numerical experiments on multiport array antenna S-parameters demonstrate the effectiveness of our proposed algorithms within the Loewner framework, where the proposed algorithms reach the smallest errors for the smallest number of frequency points chosen.

Adaptive metamodeling approach for real-time taxiing time estimation on airport surface

Investigating adaptive sampling strategies for optimal building energy performance using artificial neural networks and kriging surrogate models

An improved adaptive sampling and experiment design method for aerodynamic optimization

Demand for goods delivery services keeps growing, creating problems with last mile delivery. While autonomous delivery robots provide a solution to many of these problems, configuring their deployment to balance cost, performance, and safety is complex. Our industry partner Panasonic uses a multi-objective search approach to find configurations for their system, leveraging a simulator to assess the quality of the different configurations. However, due to the complexity of the simulator, this approach proved very computationally expensive. In this work, we propose an approach that uses a surrogate model to speed up fitness computation. Since building the surrogate model has a cost itself, the approach trains the model incrementally during the search: first, an imprecise model is used, while, as the search progresses, more precise models are used. The search stops when no improved solutions can be found. Experimental results show that this new approach, while it produces configurations of slightly lower quality, requires significantly less computing time. This is an extended abstract of "Alternating Between Surrogate Model Construction and Search for Configurations of an Autonomous Delivery System" by C.-H. Sun, T. Laurent, P. Arcaini, and F. Ishikawa, IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER 2024).

This research compares the performance of space-time surrogate models (STSMs) and network surrogate models (NSMs). Specifically, when the system response varies over time (or pseudo-time), the surrogates must predict the system response. A surrogate model is used to approximate the response of computationally expensive spatial and temporal fields resulting from some computational mechanics simulations. Within a design context, a surrogate takes a vector of design variables that describe a current design and returns an approximation of the design’s response through a pseudo-time variable. To compare various radial basis function (RBF) surrogate modeling approaches, the prediction of a load displacement path of a snap-through structure is used as an example numerical problem. This work specifically considers the scenario where analytical sensitivities are available directly from the computational mechanics’ solver and therefore gradient enhanced surrogates are constructed. In addition, the gradients are used to perform a domain transformation preprocessing step to construct surrogate models in a more isotropic domain, which is conducive to RBFs. This work demonstrates that although the gradient-based domain transformation scheme offers a significant improvement to the performance of the space-time surrogate models (STSMs), the network surrogate model (NSM) is far more robust. This research offers explanations for the improved performance of NSMs over STSMs and recommends future research to improve the performance of STSMs.

Many systems nowadays follow not only physical laws but also manmade rules. These systems are known as discrete-event dynamic systems (DEDSs), where simulation is the only faithful way for performance evaluation. Due to various advantages in practice, designs (or solution candidates) with low descriptive complexity (called simple designs) are usually preferred over complex ones when their performances are close. However, the descriptive complexity (DC) is usually nonlinear and takes discrete value, which makes traditional methods such as linear programming and gradient-based local search not applicable. Existing methods for simulation-based optimization (SBO) do not explore the preference on descriptive complexity and thus cannot solve the problem efficiently. The major contributions of this paper are to point out the importance of considering SBO problems with DC preference, and to develop an adaptive sampling algorithm (ASA) to find the simplest good design. It is shown that ASA terminates within finite iterations and with controllable probability of making mistake. The computational complexity of ASA and its dependence on various parameters are discussed. ASA is then applied to three parameter optimization problems and a node activation policy optimization problem in a wireless sensor network. Numerical results show that ASA is more efficient than blind picking and Levin search in most cases. We hope this work can shed some insight to how to find simple and good designs in general.

In order to reduce the time and resources devoted to design-space exploration during simulation-based design and optimization, the use of surrogate models, or metamodels, has been proposed in the literature. Key to the success of metamodeling efforts are the experimental design techniques used to generate the combinations of input variables at which the computer experiments are conducted. Several adaptive sampling techniques have been proposed to tailor the experimental designs to the specific application at hand, using the already-acquired data to guide further exploration of the input space, instead of using a fixed sampling scheme defined a priori. Though mixed results have been reported, it has been argued that adaptive sampling techniques can be more efficient, yielding better surrogate models with less sampling points. In this paper, we address the problem of adaptive sampling for single and multi-response metamodels, with a focus on Multi-stage Multi-response Bayesian Surrogate Models (MMBSM). We compare distance-optimal latin hypercube sampling, an entropy-based criterion and the maximum cross-validation variance criterion, originally proposed for one-dimensional output spaces and implemented in this paper for multi-dimensional output spaces. Our results indicate that, both for single and multi-response surrogate models, the entropy-based adaptive sampling approach leads to models that are more robust to the initial experimental design and at least as accurate (or better) when compared with other sampling techniques using the same number of sampling points.

Multi-fidelity neural network for uncertainty quantification of chemical reaction models

We present novel adaptive sampling algorithms for particle-based fluid simulation. We introduce a sampling condition based on geometric local feature size that allows focusing computational resources in geometrically complex regions, while reducing the number of particles deep inside the fluid or near thick flat surfaces. Further performance gains are achieved by varying the sampling density according to visual importance. In addition, we propose a novel fluid surface definition based on approximate particle-to-surface distances that are carried along with the particles and updated appropriately. The resulting surface reconstruction method has several advantages over existing methods, including stability under particle resampling and suitability for representing smooth flat surfaces. We demonstrate how our adaptive sampling and distance-based surface reconstruction algorithms lead to significant improvements in time and memory as compared to single resolution particle simulations, without significantly affecting the fluid flow behavior.

An evaluation of adaptive surrogate modeling based optimization with two benchmark problems

Surrogate models or metamodels are commonly used to exploit expensive computational simulations within a design optimization framework. The application of multi-fidelity surrogate modeling approaches has recently been gaining ground due to the potential for further reductions in simulation effort over single fidelity approaches. However, given a black box problem when exactly should a designer select a multi-fidelity approach over a single fidelity approach and vice versa? Using a series of analytical test functions and engineering design examples from the literature, the following paper illustrates the potential pitfalls of choosing one technique over the other without a careful consideration of the optimization problem at hand. These examples are then used to define and validate a set of guidelines for the creation of a multi-fidelity Kriging model. The resulting guidelines state that the different fidelity functions should be well correlated, that the amount of low fidelity data in the model should be greater than the amount of high fidelity data and that more than 10 % and less than 80 % of the total simulation budget should be spent on low fidelity simulations in order for the resulting multi-fidelity model to perform better than the equivalent costing high fidelity model.