Search results

In real life, excitations are highly non-stationary in frequency and amplitude, which easily induces resonant vibration to structural responses. Conventional control algorithms in this case cannot guarantee cost-effective control effort and efficient structural response alleviation. To this end, this paper proposes a novel adaptive linear quadratic regulator (LQR) by integrating wavelet transform and genetic algorithm (GA).

In each time interval, multiresolution analysis of real-time structural responses returns filtered time signals dominated by different frequency bands. Minimization of cost function in each frequency band obtains control law and gain matrix that depend on temporal-frequency band, so suppressing resonance-induced filtered response signal can be directly achieved by regulating gain matrix in the temporal-frequency band, leading to emphasizing cost-function weights on control and state. To efficiently subdivide gain matrices in resonant and normal frequency bands, the cost-function weights are optimized by a developed procedure associated to genetic algorithm. Single-degree-of-freedom (SDOF) and multi-degree-of-freedom (MDOF) structures subjected to near- and far-fault ground motions are studied.

Resonant band requires a larger control force than non-resonant band to decay resonance-induced peak responses. The time-varying cost-function weights generate control force more cost-effective than time-invariant ones. The scheme outperforms existing control algorithms and attains the trade-off between response suppression and control force under non-stationary excitations.

Proposed control law allocates control force amounts depending upon resonant or non-resonant band in each time interval. Cost-function weights and wavelet decomposition level are formulated in an elegant manner. Genetic algorithm-based optimization cost-efficiently results in minimizing structural responses.

This chapter examines the main methodological issues involved in the comprehension of the cost structure of the airport industry and suggests considerations for future airport cost analyses. Such understanding has become a crucial concern for policy makers, regional planners, and managers in order to deal with optimal market design (e.g., regulation and market configuration) and airport strategies (e.g., pricing, investments, and alliances). An in-depth analysis of the economics of cost functions is presented, together with a description of the relevant multi-output cost economies measures (average incremental costs, scale and scope economies, and cost complementarities). We also discuss the assumptions underlying estimates of total versus variable cost functions and the importance of estimating a sufficiently flexible functional form. Moreover, we provide a critical survey of the international empirical literature on the cost structure of the airport industry, which highlights how econometric estimates strongly depend on the sample choice and the empirical model considered. Indeed, while econometric studies on international samples based on long-run cost function estimates show that long-run scale economies are never exhausted, single country studies mostly estimate variable cost functions and find lower values for scale economies at median sample points that tend to decrease with size. We discuss why we believe that studies based on the estimation of short-run variable cost functions offer more reliable results, given the reasonable assumption of airport overcapitalization in the short run. We conclude our work by noting that underlying policy issues related to planning and regulation, as well as to the optimal market structure of the airport sector, need to take into account the role played by vertical relationships between airports and airlines.

Contemporary stochastic optimal control by synergy of the probability density evolution method (PDEM) and conventional optimal controller exhibits less capability to guarantee economical energy consumption versus control efficacy when non-stationary stochastic excitations drive hysteretic structures. In this regard, a novel multiscale stochastic optimal controller is invented based on the wavelet transform and the PDEM.

For a representative point, a conventional control law is decomposed into sub-control laws by deploying the multiresolution analysis. Then, the sub-control laws are classified into two generic control laws using resonant and non-resonant bands. Both frequency bands are established by employing actual natural frequency(ies) of structure, making computed efforts depend on actual structural properties and time-frequency effect of non-stationary stochastic excitations. Gain matrices in both bands are then acquired by a probabilistic criterion pertaining to system second-order statistics assessment. A multi-degree-of-freedom hysteretic structure driven by non-stationary and non-Gaussian stochastic ground accelerations is numerically studied, in which three distortion scenarios describing uncertainties in structural properties are considered.

Time-frequency-dependent gain matrices sophisticatedly address non-stationary stochastic excitations, providing efficient ways to independently suppress vibrations between resonant and non-resonant bands. Wavelet level, natural frequency(ies), and ratio of control forces in both bands influence the scheme’s outcomes. Presented approach outperforms existing approach in ensuring trade-off under uncertainty and randomness in system and excitations.

Presented control law generates control efforts relying upon resonant and non-resonant bands, and deploys actual structural properties. Cost-function weights and probabilistic criterion are promisingly developed, achieving cost-effectiveness of energy demand versus controlled structural performance.

Improving the accuracy and reducing computational costs of predictions, especially the prediction of time series, is one of the most critical parts of the decision-making processes and management in different areas and organizations. One of the best solutions to achieve high accuracy and low computational costs in time series forecasting is to develop and use efficient hybrid methods. Among the combined methods, parallel hybrid approaches are more welcomed by scholars and often have better performance than sequence ones. However, the necessary condition of using parallel combinational approaches is to estimate the appropriate weight of components. This weighting stage of parallel hybrid models is the most effective factor in forecasting accuracy as well as computational costs. In the literature, meta-heuristic algorithms have often been applied to weight components of parallel hybrid models. However, such that algorithms, despite all unique advantages, have two serious disadvantages of local optima and iterative time-consuming optimization processes. The purpose of this paper is to develop a linear optimal weighting estimator (LOWE) algorithm for finding the desired weight of components in the global non-iterative universal manner.

In this paper, a LOWE algorithm is developed to find the desired weight of components in the global non-iterative universal manner.

Empirical results indicate that the accuracy of the LOWE-based parallel hybrid model is significantly better than meta-heuristic and simple average (SA) based models. The proposed weighting approach can improve 13/96%, 11/64%, 9/35%, 25/05% the performance of the differential evolution (DE), genetic algorithm (GA), particle swarm optimization (PSO) and SA-based parallel hybrid models in electricity load forecasting. While, its computational costs are considerably lower than GA, PSO and DE-based parallel hybrid models. Therefore, it can be considered as an appropriate and effective alternative weighing technique for efficient parallel hybridization for time series forecasting.

In this paper, a LOWE algorithm is developed to find the desired weight of components in the global non-iterative universal manner. Although it can be generally demonstrated that the performance of the proposed weighting technique will not be worse than the meta-heuristic algorithm, its performance is also practically evaluated in real-world data sets.

Flying‐capacitor multilevel converters (FCC) need a passive or active regulation of the capacitor voltages. Recently the trend is towards active control, often implemented separately from the current control. The advantages of a true multi‐variable control sparked the interest to apply Model Based Predictive Control (MBPC) for FCC. In this paper an objective analysis method to evaluate the effects of several design choices is presented. The effects of the weight factor selection, model simplification, and prediction horizon expansion for MBPC of a 3‐level FCC are analyzed in a systematical way.

The analysis is mainly based on the mean square error (MSE) of current and capacitor voltage. The results are analysed for different lengths of the prediction horizon and for a wide range of weight factor values. Similarly the effect of a model simplification, neglecting the neutral point voltage, is studied when implementing MBPC for FCCs while considering the computational aspects. Validation of the simulation results is done by experiments on an FPGA‐based setup.

Including the effect of the neutral point voltage considerably increases the current control quality and a much wider range of good values for the weight factor exists. As this good range is not critically dependent on the current amplitude it is possible to select one weight factor value for all operating points. Furthermore, it is concluded that increasing the prediction horizon increases the computational load without improving the control quality.

The effects of increasing the prediction horizon when including other controlled variables is to be investigated, as well as the robustness to modeling errors. The MSE analysis methodology is very suitable for this further research.

For practitioners of MBPC in power electronics the paper proves that by means of simulations and the MSE one value for weight factor can be chosen for all operating points. The paper clearly shows that a practical implementation is feasible and demonstrates that neglecting the neutral point voltage is not good practice.

The MSE‐based analysis is shown to be a systematical and unbiased methodology to evaluate the effects of design choices. The results from this analysis can be directly applied in practical setups.

The purpose of this paper is to describe a methodology for characterization of the robot environment to help solve such problem as designing an optimal agricultural robot for a specific agricultural task.

Defining and characterizing a task is a crucial step in the optimization of a task-specific robot. It is especially difficult in the agricultural domain because of the complexity and unstructured nature of the environment. In this research, trees are modeled from orchards and are used as the robot working environment, the geometrical features of an agricultural task are investigated and a method for designing an optimal agricultural robot is developed. Using this method, a simplified characteristic environment, representing the actual environment, is developed and used.

Case studies showing that the optimal robot, which is designed based on the characteristic environment, is similar to the optimal robot, which is designed based on the actual environment (less than 4 per cent error), is presented, while the optimization run time is significantly shorter (up to 22 times) when using the characteristic environment.

This paper proposes a new concept for solving the robot task-based optimization by the analysis of the task environment and characterizing it by a simpler artificial task environment. The methodology decreases the time of the optimal robot design, allowing to take into account more details in an acceptable time.

Maintaining a balanced neutral point, reducing power loss, execution time are important criteria for the controlling of neutral point clamped (NPC) inverter. However, it is tough to meet all the challenges and also supplying the load current within the harmonic limit. This paper aims to maintain load current quality within the Institute of Electrical and Electronics Engineers 519 standard and meet the above-mentioned challenges.

The output load current of a three-level simplified neutral point clamped (3 L-SNPC) inverter is controlled in this paper using model predictive control (MPC). The 3 L-SNPC inverters is considered because fewer semiconductor devices are used in this topology; this will enhance the reliability of the system. MPC is used as a controller because it can handle the direct current-link capacitors’ voltage balancing problem in a very intuitive way. The proposed 3 L-SNPC yields similar current total harmonic distortion (THD), transient and steady-state responses, voltage stress and over current protection capability as the conventional NPC inverter. To reduce the computational burden of the proposed SNPC system, two simplified MPC strategies are proposed, namely, single voltage vector prediction-based MPC and selective voltage vector prediction-based MPC.

The system shows a current THD of 2.33% at 8.96 kHz. The overall loss of the system is reduced significantly to be useful in medium power applications. The required execution times for the simplified MPC strategies are tested on the hardware dSPACE 1104 platform. It is found that the single voltage vector prediction-based MPC and the selective voltage vector prediction-based MPC are computationally efficient by 8.28% and 62.9%, respectively, in comparison with the conventional MPC-based conventional NPC system.

Multiple system constraints are considered throughout the paper and also compare the SNPC to the conventional NPC inverter. Proper current tracking, over-current protection, overall power loss reduction especially switching loss and maintaining capacitor voltages balance at a neutral point are achieved. The improvement of execution time has also been verified and calculated using hardware-in-loop of the dSPACE DS1104 platform.

Community detection of dynamic networks provides more effective information than static network community detection in the real world. The mainstream method for community detection in dynamic networks is evolutionary clustering, which uses temporal smoothness of community structures to connect snapshots of networks in adjacent time intervals. However, the error accumulation issues limit the effectiveness of evolutionary clustering. While the multi-objective evolutionary approach can solve the issue of fixed settings of the two objective function weight parameters in the evolutionary clustering framework, the traditional multi-objective evolutionary approach lacks self-adaptability.

This paper proposes a community detection algorithm that integrates evolutionary clustering and decomposition-based multi-objective optimization methods. In this approach, a benchmark correction procedure is added to the evolutionary clustering framework to prevent the division results from drifting.

Experimental results demonstrate the superior accuracy of this method compared to similar algorithms in both real and synthetic dynamic datasets.

To enhance the clustering results, adaptive variances and crossover probabilities are designed based on the relative change amounts of the subproblems decomposed by MOEA/D (A Multiobjective Optimization Evolutionary Algorithm based on Decomposition) to dynamically adjust the focus of different evolutionary stages.