Explore the vast range of titles available.

Convolutional Neural Network (CNN) is a branch of deep learning which has been one of a popular methods in different applications, especially in medical field. In this study, an optimized CNN model is built using the random search optimization to classify seven types of skin cancer, namely, basal cell carcinoma, melanoma, dermatofibroma, vascular lesion, melanocytic nevus, actinic keratosis and benign keratosis. Total of 10,015 images were collected from the Human Against Machine dataset (HAM10000) which is available in Kaggle, Even though CNN has shown best results in many applications, the hyper-parameters that are required to build CNN model is difficult to choose. If the chosen hyper-parameters doesn’t show good results, the model should be trained again with other set of hyper-parameter values. To avoid this circumstance, the hyper-parameter optimization is required and in this study, it is done using random search optimization. A base CNN model is initially created without using any optimization technique, so that the performance of the CNN model which is optimized by the random search method can be compared and analysed. The first model provided an accuracy of 73.34%, whereas the optimized model shown an improvement in accuracy of 77.17%.

A modified version of a common global optimization method named controlled random search is presented here. This method is designed to estimate the global minimum of multidimensional symmetric and asymmetric functional problems. The new method modifies the original algorithm by incorporating a new sampling method, a new termination rule and the periodical application of a local search optimization algorithm to the points sampled. The new version is compared against the original using some benchmark functions from the relevant literature.

Since nature is an excellent source of inspiration for optimization methods, many optimization algorithms have been proposed, are inspired by nature, and are modified to solve various optimization problems. This paper uses metaheuristics in a new field inspired by nature; more precisely, we use pollination optimization in cocoa plants. The cocoa plant was chosen as the object since its flower type differs from other kinds of flowers, for example, by using cross-pollination. This complex relationship between plants and pollinators also renders pollination a real-world problem for chocolate production. Therefore, this study first identified the underlying optimization problem as a deferred fitness problem, where the quality of a potential solution cannot be immediately determined. Then, the study investigates how metaheuristic algorithms derived from three well-known techniques perform when applied to the flower pollination problem. The three techniques examined here are Swarm Intelligence Algorithms, Individual Random Search, and Multi-Agent Systems search. We then compare the behavior of these various search methods based on the results of pollination simulations. The criteria are the number of pollinated flowers for the trees and the amount and fairness of nectar pickup for the pollinator. Our results show that Multi-Agent System performs notably better than other methods. The result of this study are insights into the co-evolution of behaviors for the collaborative pollination task. We also foresee that this investigation can also help farmers increase chocolate production by developing methods to attract and promote pollinators.

We propose an optimization-via-simulation algorithm, called COMPASS, for use when the performance measure is estimated via a stochastic, discrete-event simulation, and the decision variables are integer ordered. We prove that COMPASS converges to the set of local optimal solutions with probability 1 for both terminating and steady-state simulation, and for both fully constrained problems and partially constrained or unconstrained problems under mild conditions.

Practical optimization problems are often too complex to be formulated exactly. Knowing multiple good alternatives can help decision-makers easily switch solutions when needed, such as when faced with unforeseen constraints. A multimodal optimization task aims to find multiple global optima as well as high-quality local optima of an optimization problem. Evolutionary algorithms with niching techniques are commonly used for such problems, where a rough estimate of the optima number is required to determine the population size. In this paper, a partition-based random search method is proposed, in which the entire feasible domain is partitioned into smaller and smaller subregions iteratively. Promising regions are partitioned faster than unpromising regions, thus, promising areas will be exploited earlier than unpromising areas. All promising areas are exploited in parallel, which allows multiple good solutions to be found in a single run. The proposed method does not require prior knowledge about the optima number and it is not sensitive to the distance parameter. By cooperating with local search to refine the obtained solutions, the proposed method demonstrates good performance in many benchmark functions with multiple global optima. In addition, in problems with numerous local optima, high-quality local optima are captured earlier than low-quality local optima.

* There was significant absorption of heavy metals by the pepper in contaminated soils. * The target hazard quotient (THQ) indices followed the order of Pb>Zn>>Cd » Ni. * Relationships exist between contaminated plants and electromagnetic wave. * PCA and random search can select the main spectra and predict THQ for each element. Given the tendency of heavy metals to accumulate in soil and plants, the purpose of this study was to determine the contamination levels of Cd, Ni, Pb, and Zn on peppers (leaves and fruit) grown in contaminated soils in industrial centers. For this purpose, we measured the uptake of the four heavy metals by peppers grown in the heavy metal contaminated soils throughout the four growth stages: two-leaf, growth, flowering, and fruiting, and calculated various vegetation indices to evaluate the heavy metal contamination potentials. Electromagnetic waves were also applied for analyzing the responses of the target plants to various heavy metals. Based on the relevant spectral bands identified by principal component analysis (PCA) and random search methods, a regression method was then employed to determine the most optimal spectral bands for estimating the target hazard quotient (THQ). The THQ was found to be the highest in the plants contaminated by Pb (THQ= 62) and Zn (THQ= 5.07). The results of PCA and random search indicated that the spectra at the bands of b 570 , b 650 , and b 760 for Pb, b 400 and b 1030 for Ni, b 400 and b 880 for Cd, and b 560 , b 910 , and b 1050 for Zn were the most optimal spectra for assessing THQ. Therefore, in future studies, instead of examining the amount of heavy metals in plants by chemical analysis in the laboratory, the responses of the plants to the electromagnetic waves in the identified bands can be readily investigated in the field based on the established correlations.

In this paper, we introduce a novel iterated random search method for large-scale texture synthesis and manipulations. Previous researches on texture synthesis and manipulation have reached a great achievement both on quality and performance. However, the cost of the popular exhaustive search-based methods is still high especially for large-scale and complex synthesis scenes. Our algorithm contributes great improvements on performances about 2–50 times over the typical patch-based synthesis methods. Texture patterns have been well-known formalized as a Markov Random Field (MRF) whose two hypotheses, stationarity and locality, drive our bold guess that a random sampling may just catch a good match and allows us to propagate the natural coherence in the neighborhood. Meanwhile, the iteration constantly updates the bad guesses to make our algorithm converge fast with the results in the state of the art. We also provide a simple theoretical analysis to compare our iterated randomized search model and the classical synthesis algorithms. Besides, this simple method turns out to work well in various applications as well, such as texture transfer, image completion and video synthesis.

As we all know, multi-level imperfect maintenance strategy is usually more effective than single-level maintenance strategy for the actual production machines. At the same time, in the previous multi-level maintenance strategies, the majority of maintenance models only consider the constant maintenance thresholds, while variable maintenance thresholds are usually ignored. Under these contexts, a novel multi-level imperfect maintenance model with variable preventive maintenance (PM) thresholds, variable overhaul maintenance (OM) thresholds and variable number of PMs in each OM cycle is established. In order to deal with the concerned problem, a novel twin delayed deep deterministic policy gradient (TD3) algorithm that is a kind of reinforcement learning is designed, renamed as NTD3. Finally, through numerical simulation, we can find that (1) the average improvements between the proposed maintenance strategy and other three traditional strategies in the average cost rate (ACR) are 11.50%, 595.91% and 5.16%, respectively; and (2) the average improvement between the proposed NTD3 and other random search method is 5.53%. Thus, the effectiveness of the proposed maintenance strategy and the superiority of proposed NTD3 are all demonstrated.

SVM with an RBF kernel is usually one of the best classification algorithms for most data sets, but it is important to tune the two hyperparameters C and γ to the data itself. In general, the selection of the hyperparameters is a non-convex optimization problem and thus many algorithms have been proposed to solve it, among them: grid search, random search, Bayesian optimization, simulated annealing, particle swarm optimization, Nelder Mead, and others. There have also been proposals to decouple the selection of γ and C. We empirically compare 18 of these proposed search algorithms (with different parameterizations for a total of 47 combinations) on 115 real-life binary data sets. We find (among other things) that trees of Parzen estimators and particle swarm optimization select better hyperparameters with only a slight increase in computation time with respect to a grid search with the same number of evaluations. We also find that spending too much computational effort searching the hyperparameters will not likely result in better performance for future data and that there are no significant differences among the different procedures to select the best set of hyperparameters when more than one is found by the search algorithms.

Electric energy, as an economical and clean energy, plays a significant role in the development of science and technology and the economy. The motor is the core equipment of the power station; therefore, monitoring the motor vibration and predicting time series of the bearing vibration can effectively avoid hazards such as bearing heating and reduce energy consumption. Time series forecasting methods of motor bearing vibration based on sliding window forecasting, such as CNN, LSTM, etc., have the problem of error accumulation, and the longer the time-series forecasting, the larger the error. In order to solve the problem of error accumulation caused by the conventional methods of time series forecasting of motor bearing vibration, this paper innovatively introduces Informer into time series forecasting of motor bearing vibration. Based on Transformer, Informer introduces ProbSparse self-attention and self-attention distilling, and applies random search to optimize the model parameters to reduce the error accumulation in forecasting, achieve the optimization of time and space complexity and improve the model forecasting. Comparing the forecasting results of Informer and those of other forecasting models in three publicly available datasets, it is verified that Informer has excellent performance in time series forecasting of motor bearing vibration and the forecasting results reach 10−2∼10−6.