Crop evapotranspiration (ETC) is an important component of the agricultural hydrological cycle, and its accurate estimation is important for assessing the hydrological environment of farmlands and guiding crop irrigat...
详细信息
Crop evapotranspiration (ETC) is an important component of the agricultural hydrological cycle, and its accurate estimation is important for assessing the hydrological environment of farmlands and guiding crop irrigation. To improve the prediction accuracy of ETC for early and late rice, this study analyzed the importance of the influencing factors based on a categorical regression tree (CART) machine learning algorithm. Based on the importance results of the influencing factors obtained from CART analysis, different combinations of input factors were constructed, and daily ETC prediction models were developed for early and late rice using the least squares support vector machine (LSSVM). A three-element heuristicoptimization algorithm, Sine Cosine Algorithm (SCA), Slime Mold Algorithm (SMA), and Sparrow Search Algorithm (SSA), were used to optimize the ETC model to improve the performance of the early and late rice ETC prediction models. The results of this study revealed that the strongest correlation between early and late rice ETC was net solar radiation (Rn), with importance of 0.830 and 0.802, respectively. The next strongest correlations were temperature (T), wind speed (u), soil moisture (SM) and u, leaf area index (LAI), SM (0.092, 0.02, 0.014 and 0.058, 0.049, 0.034), respectively. The accuracy of the model tended to increase as the number of influencing factors in the input combinations constructed based on the importance results increased. The accuracy of the model increased more rapidly when the input factors were one to four and the accuracy of the model increased slightly when the input factors were more than four. The SSA-LSSVM model with four-factor combinations of inputs was the least-factor combination with guaranteed model accuracy of RMSE = 0.514 mm/d, R2 = 0.924, KGE = 0.565, WI = 0.978,and MAE = 0.414 mm/d;RMSE = 0.468 mm/d, R2 = 0.910, KGE = 0.670, WI = 0.985, and MAE = 0.368 mm/d. This study may provide a reference for the prediction of eva
Automatic image segmentation is a challenging task in computer vision applications, especially in the presence of occluded objects, varying color, and different lighting conditions. The advancement of depth-sensing te...
详细信息
Automatic image segmentation is a challenging task in computer vision applications, especially in the presence of occluded objects, varying color, and different lighting conditions. The advancement of depth-sensing technologies has introduced RGB-Depth cameras which are capable to generate RGB-Depth images and brought significant changes in computer vision applications. However, the segmentation of RGB-Depth images is a difficult task. Therefore, in this paper, a new segmentation method for RGB-Depth images has been introduced and named as random Henry gas solubility optimization-fuzzy clustering method. Firstly, a random Henry gas solubility optimization algorithm has been developed. Next, the proposed optimization algorithm has been employed to obtain optimal fuzzy clusters which are finally merged through segmentation by aggregating superpixels. The standard NYU depth V2 RGB-Depth indoor image dataset is used for performance evaluation. The proposed segmentation approach has been compared with five different methods namely, kmeans, fuzzy c-means, Henry gas solubility optimization algorithm, chaotic gravitational search algorithm, and J Segmentation in terms of qualitative and quantitative measures. The result analysis shows that the proposed RGB-D segmentation method outperforms the other considered methods.
Swarm intelligence meta-heuristic optimization algorithms for optimizing engineering applications have become increasingly popular. The whale optimization algorithm (WOA) is a recent and effective swarm intelligence o...
详细信息
Swarm intelligence meta-heuristic optimization algorithms for optimizing engineering applications have become increasingly popular. The whale optimization algorithm (WOA) is a recent and effective swarm intelligence optimization algorithm that mimics humpback whales' behaviors when optimizing a problem. Applying the algorithm to achieve optimal solutions has shown good results compared to most meta-heuristic optimization algorithms. However, complex applications might require the processing of large-scale computations, which results in down-scaling computational throughput of WOA. Apache Spark, a well-known parallel data processing framework, is the most recent distributed computing framework and has been proven to be the most efficient. In this article, we propose a WOA implementation on top of Apache Spark, represented as SBWOA, to enhance its computational performance while providing higher scalability of the algorithm for handling more complex problems. Compared with the recently reported MapReduce WOA (MR-WOA), and serial implementation of WOA, our approach achieves significant enhancements with respect to computational performance for the highest population size with the maximum number of iterations. SBWOA successfully handles higher-complexity problems which require complex computations.
The current work presents a comparative study of hybrid models that use support vector machines (SVMs) and meta-heuristic optimization algorithms (MOAs) to predict the ultimate interfacial bond strength (IBS) capacity...
详细信息
The current work presents a comparative study of hybrid models that use support vector machines (SVMs) and meta-heuristic optimization algorithms (MOAs) to predict the ultimate interfacial bond strength (IBS) capacity of fiber-reinforced polymer (FRP). More precisely, a dataset containing 136 experimental tests was first collected from the available literature for the development of hybrid SVM models. Five MOAs, namely the particle swarm optimization, the grey wolf optimizer, the equilibrium optimizer, the Harris hawks optimization and the slime mold algorithm, were used;five hybrid SVMs were constructed. The performance of the developed SVMs was then evaluated. The accuracy of the constructed hybrid models was found to be on the higher side, with R-2 ranges between 0.8870 and 0.9774 in the training phase and between 0.8270 and 0.9294 in the testing phase. Based on the experimental results, the developed SVM-HHO (a hybrid model that uses an SVM and the Harris hawks optimization) was overall the most accurate model, with R-2 values of 0.9241 and 0.9241 in the training and testing phases, respectively. Experimental results also demonstrate that the developed hybrid SVM can be used as an alternate tool for estimating the ultimate IBS capacity of FRP concrete in civil engineering projects.
Harris hawks optimization (HHO) algorithm inspired by the cooperative behavior and chasing style of harris' hawks in nature called surprise pounce is a relatively novel swarm intelligent optimization algorithm. Du...
详细信息
Harris hawks optimization (HHO) algorithm inspired by the cooperative behavior and chasing style of harris' hawks in nature called surprise pounce is a relatively novel swarm intelligent optimization algorithm. Due to its simplicity and efficiency, the canonical HHO has displayed promising performance on a large number of continuous optimization problems and real-world optimization problems. However, how to balance contradictions between the exploration and exploitation capabilities and alleviate the premature convergence are two critical concerns that need to be dealt with in the HHO study. To address these two drawbacks, improve the optimization performance, and broaden its application domain, a dynamic multi-swarm differential learning Harris hawks optimizer (DMSDLHHO) is proposed in this paper. To efficiently maintain the population diversity, the whole population is divided into many small sub-swarms, which are regrouped periodically, and information is exchanged among the swarms. In each generation, the differential evolution operator (including mutation, crossover, and selection operators) based on the personal historical best position is merged into each sub-swarm to augment the exploration capability, while the Quasi-Newton method as a local searcher is used to enhance the exploitation capability. Besides, aiming to prevent the algorithm from falling into local optima to some extent, the differential mutation operator candidate pool strategy is introduced into the late stage of the search process. Thus, different individuals in the same population can conduct distinct search behaviors in each generation, and the same individual can perform various search behaviors in different generations. The proposed algorithm is tested on 23 classic test functions and 30 CEC2014 benchmark functions and is compared with quite a few state-of-the-art algorithms in terms of often-used performance metrics with the help of statistical analysis, diversity measurement, explorati
RFID technology is increasingly incorporated in many facets of life and accordingly represents an active research area. RFID network planning is a hard optimization problem that determines positions of readers and tra...
详细信息
ISBN:
(纸本)9781509051601
RFID technology is increasingly incorporated in many facets of life and accordingly represents an active research area. RFID network planning is a hard optimization problem that determines positions of readers and transmitter power parameters in order to satisfy requirements about coverage, interference, power consumption, total cost, etc. For such hard optimization problems, where deterministic mathematical methods are inadequate, stochastic swarm intelligence algorithms are very successful. In this paper we adjusted the recent guided fireworks algorithm for the RFID network planning problem with probabilistic model of coverage. Our proposed approach was tested on standard benchmark networks and compared with other algorithms from literature. Our approach proved to be better than other compared methods, considering the coverage, number of employed readers, used power and interference.
This paper employs a deep learning network with a comprehensive architecture to forecast Bitcoin prices, enhancing accuracy by integrating two meta-heuristic optimization algorithms, INFO and NRBO. Empirical results d...
详细信息
This paper employs a deep learning network with a comprehensive architecture to forecast Bitcoin prices, enhancing accuracy by integrating two meta-heuristic optimization algorithms, INFO and NRBO. Empirical results demonstrate that the hybrid model significantly outperforms the LSTM in both fit and predictive accuracy across in-sample and out-of-sample data. Notably, the NRBO-CNN-BiLSTM-Attention model substantially improves accuracy in 5-day and 15-day forecasts, reducing the MAPE by over 50 % compared to the LSTM model, thereby significantly enhancing overall predictive performance. The robustness of our results is supported by the MCS tests. Furthermore, strategically modifying time steps in data analysis optimizes model performance.
暂无评论