Optimization and Engineering最新文献

To avoid the problem of alternating layers of different materials in the thickness aspect, a topology optimization method of composite laminate considering density change rate constraint is proposed. This method utilizes the density of a specific layer to constrain the upper limit of density for its neighboring layers, so that the relative density of the upper and lower layers is greater or less than the middle layers. The middle layers of the laminate are one material and the adjacent upper and lower layers are another material. The low-density material in the middle layers is taken as an example, the density of the specified layer in the design space is used to constrain the upper limit of the density of its adjacent layers. The middle layers are limited by the constraint strategy, and the relative density is smaller than that of the two sides. The purpose of replacing the middle layer where is in the design domain with low-density material can be effectively realized. The mathematical model for patch topology optimization of composite laminate considering density change rate constraint is established, and the reasonable space layout of fiber composite and low-density material is obtained by solving. The numerical example of the composite laminate and the wind turbine blade structure show that the optimized two-phase materials distribution follows the corresponding manufacturing constraints, and also reduces the total mass of the structure while ensuring the mechanical properties. And the mass of their structures are reduced while ensuring the mechanical properties. The feasibility and effectiveness of the method are verified.

We propose an effective algorithm for black-box optimization problems without derivatives in the presence of output constraints. The proposed algorithm is illustrated using a realistic short-term oil production case with complex functions describing system dynamics and output constraints. The results show that our algorithm provides feasible and locally near-optimal solutions for a complex decision-making problem under uncertainty. The proposed algorithm relies on building approximation models using a reduced number of function evaluations, resulting from (i) an efficient model improvement algorithm, (ii) a decomposition of the network of wells, and (iii) using a spectral method for handling uncertainty. We show, in our case study, that the use of the approximation models introduced in this paper can reduce the required number of simulation runs by a factor of 40 and the computation time by a factor of 2600 compared to the Monte Carlo simulation with similarly satisfactory results.

Battery electric vehicles with multiple motors are characterized by actuator redundancy, which calls for application of instantaneously optimized distribution of motor/wheel torques, thus minimizing the energy consumption, i.e., maximizing the vehicle range. If the e-motors are equipped with disconnect clutches, the energy saving potential becomes even higher due to the avoidance of drag of inactive electric motors. However, in this case optimization through time and predictive control techniques should be used to provide globally minimal energy consumption. To this end, the paper proposes the following modeling, optimization, and model predictive control method for straight-line driving mode: (i) a dynamic backward-looking model of electric vehicle propelled by disconnect clutch-equipped four wheel motors, which takes into account the clutch synchronization-related drivetrain transient loss; (ii) globally optimal, dynamic programming (DP)-based off-line optimization of e-motor torque and clutch state control trajectories, and (iii) a model predictive torque vectoring control (MPC) strategy. The MPC strategy is verified by simulation for various certification driving cycles, and the results are compared with the DP-optimal benchmark for different values of a user-defined weighting coefficient, which penalizes frequent clutch disconnects for improved durability. The DP optimization results reveal that the energy consumption reduction achieved through the disconnect clutch functionality is up to 7%, on top of up to 5% reduction achieved by torque distribution itself. The MPC control strategy relying on the prediction horizon of 10 steps approach the DP energy consumption benchmark within the margin of 1%.

Automotive noise source identification is important for improving driving comfort and protecting people’s auditory health. However, the stable, accurate and fast identification of low-frequency target sound sources has always been a difficult problem in the field of automotive noise source identification and sound field reconstruction. To this end, a new sound field reconstruction method, ℓ₁-Cauchy plus, is proposed in this paper, which firstly utilizes the WBH method to derive the target equivalent source strength, which is then used as the initial value for the iteration, and solved by applying the ℓ₁-Cauchy sound field reconstruction method. This hybridization process endows the proposed method with better amplitude reconstruction and improves the reconstruction of the source signal, enabling it to reconstruct the target source more efficiently in low-frequency environments. The experimental results show that the proposed method is able to accurately reconstruct the low-frequency target sound source, which is of practical application value for automobile noise control and other fields.

Continuous increase in society’s prosperity causes overwhelming growth of the produced municipal solid waste. Circular economy initiatives help to solve this problem by creating closed production cycles, where the produced waste is recycled, or its energy is recovered. An embedment of such principles requires implementation of new waste management strategies. However, these novel strategies must be based on the accurate forecasts of future waste flows. Municipal solid waste production data demonstrate behavior of hierarchical time series. Among all possible approaches to hierarchical times series forecasting, this article is focused on the reconciliation of the base waste generation forecasts. The novel method, that is based on the game-theoretically optimal reconciliation of hierarchical time series, is presented. The modified approach enables to incorporate interdependencies between time series using correlation matrix and to obtain the forecasts corresponding to the unique solution of the optimization problem. The potential of the proposed abstract approach is demonstrated on the waste production data of paper, plastics (both primarily sorted by households), and mixed municipal solid waste from the Czech Republic.

Expectile has recently gained an admiration in the area of portfolio optimization (PO) mainly because of its unique property of being both coherent and elicitable function. Additionally, a PO model minimizing Expectile function as risk measure is a linear program under discrete time setting. With these favorable features, we aim to study and analyze the Expectile and its deviation counterpart, deviation Expectile (DExpectile) based PO models in comparison to much more popular PO models comprising Conditional Value-at-Risk (CVaR) and deviation CVaR (DCVaR). We first conduct sensitivity analysis of Expectile and DExpectile PO models with respect to their two model parameters, risk-return trade-off parameter and tail-risk trade-off parameter. Thereafter, we conduct a computational analysis among Expectile, DExpectile, CVaR, and DCVaR PO models on the basis of several performance indices. Empirical study of this paper is carried out over the sample data of S &P 500 (USA) with a sample period from 06 January 2015 to 07 June 2022. Numerical results show the favorable outcomes of Expectile PO model in comparison to the models DExpectile and DCVaR, whereas it performs better than CVaR model for many likely scenarios of model parameters. On many occasions, the model DExpectile dominates DCVaR in terms of mean return, risk measures, and financial ratios while it able to outperform model CVaR under some special cases of parameters. Therefore, our numerical findings hint that the Expectile based PO models can become potential competitors to CVaR based PO models in practice.

This paper dedicates to presenting an uncertainty analysis framework for robust topology optimization (RTO) based on truss-like material model that integrates non-intrusive polynomial chaos expansion (PCE) approach. In this framework, the RTO problem is formulated as an optimization problem, which aims at minimizing both the expectancy and the standard derivation of the structural compliance with volume constraints. The magnitude and direction of load uncertainty are assumed to follow a Gaussian distribution independently. A standard non-intrusive PCE requires a large number of multivariate integrals to calculate the expansion coefficient. Therefore, response metrics such as structural compliance are efficiently characterized using the decoupling techniques based on the expansions of the uncertainty parameters. The mechanical analysis and uncertainty analysis are separated, so that the number of simulations in the original PCE procedure is greatly reduced for linear structures by means of superposition. The optimization is achieved by gradient-based methods. The appreciable accuracy and efficiency are validated by the Monte Carlo simulation. Three numerical examples are provided to demonstrate that the proposed method can lead to designs with completely different topologies and superior robustness compared to standard one.

We propose a model-based reinforcement learning method using Monte Carlo Tree Search planning. The approach assumes a black-box approximated model of the environment developed by an expert using any kind of modeling framework and it improves the model as new information from the environment is collected. This is crucial in real-world applications, since having a complete knowledge of complex environments is impractical. The expert’s model is first translated into a neural network and then it is updated periodically using data, i.e., state-action-next-state triplets, collected from the real environment. We propose three different methods to integrate data acquired from the environment with prior knowledge provided by the expert and we evaluate our approach on a domain concerning air quality and thermal comfort control in smart buildings. We compare the three proposed versions with standard Monte Carlo Tree Search planning using the expert’s model (without adaptation), Proximal Policy Optimization (a popular model-free DRL approach) and Stochastic Lower Bounds Optimization (a popular model-based DRL approach). Results show that our approach achieves the best results, outperforming all analyzed competitors.

Climate change compels a reduction of greenhouse gas emissions, yet vehicular traffic still contributes significantly to the emission of air pollutants. Hence, in this paper we focus on the optimization of traffic flow while simultaneously minimizing air pollution using speed limits as controllable parameters. We introduce a framework of traffic emission models to simulate the traffic dynamic as well as the production and spread of air pollutants. We formulate a multi-objective optimization problem for the optimization of multiple aspects of vehicular traffic. The results show that multi-objective optimization can be a valuable tool in traffic emission modeling as it allows to find optimal compromises between ecological and economic objectives.