Soft Computing最新文献

Digitizing ancient manuscripts and making them accessible to a broader audience is a crucial step in unlocking the wealth of information they hold. However, automatic recognition of handwritten text and the extraction of relevant information such as named entities from these manuscripts are among the most difficult research topics, due to several factors such as poor quality of manuscripts, complex background, presence of ink stains, cursive handwriting, etc. To meet these challenges, we propose two systems, the first system performs the task of handwritten text recognition (HTR) in ancient manuscripts; it starts with a preprocessing operation. Then, a convolutional neural network (CNN) is used to extract the features of each input image. Finally, a recurrent neural network (RNN) which has Long Short-Term Memory (LSTM) blocks with the Connectionist Temporal Classification (CTC) layer will predict the text contained in the image. The second system focuses on recognizing named entities and deciphering the relationships among words directly from images of old manuscripts, bypassing the need for an intermediate text transcription step. Like the previous system, this second system starts with a preprocessing step. Then the data augmentation technique is used to increase the training dataset. After that, the extraction of the most relevant features is done automatically using a CNN model. Finally, the recognition of names entities and the relationship between word images is performed using a bidirectional LSTM. Extensive experiments on the ESPOSALLES dataset demonstrate that the proposed systems achieve the state-of-the-art performance exceeding existing systems.

Recently, it has been observed that the weather is changing constantly because of global warming. The government is urging everyone, including scientists and the general public, to help address the severe challenges caused by climate change. Addressing the pivotal issue of carbon emissions stemming from transportation, this manuscript delves into the development of an efficient and coordinated management system. The proposed solution involves a green solid transportation system employing a two-stage network to implement a carbon cap and trade policy. A mathematical model is introduced to underscore the significance of this approach. Because of market fluctuations, supply and demand constraints are not always the same. Therefore, a two-folded uncertainty is included in this article for a better realistic outcome. A ranking defuzzification approach is employed to convert this uncertainty into a deterministic measure. Two illustrative numerical case studies are presented to underscore the effectiveness and feasibility of the proposed approaches. Then, three multi-objective techniques are employed to obtain Pareto-optimal solutions for the addressed problem. After that, a comparative study among these techniques is introduced and a sensitivity analysis is added to explore how the objective functions are influenced by potential changes in supply and demand. In conclusion, the paper offers important insights and identifies areas for future research in this field.

In this study, we propose a model called LFSPN, which serves as an extension of stochastic Petri nets dedicated to the multi-agent systems paradigm. The main objective is to specify, verify, validate, and evaluate the flow of materials within an automated production chain. We illustrate the practicality of our model by engaging in a systematic process of modeling and simulation of a production chain involving material flow. To evaluate the performance, we employ a mobile learning agent, which has distinct characteristics, namely mobility and learning. So, the distinctive characteristics of the learning agent are manifested in two key behaviors: mobility and learning. Notably, the learning agent is equipped with a flexible learning algorithm that integrates stochastic elements based on transitions. We suggest using a MATLAB simulation to determine the firing time of each transition within a sequence, guided by three different probability laws (exponential, normal, and log-normal). This sequence is designed to optimize the production process objective while facilitating learning cycles through agent rewards, specified by a production and consumption of tokens in our evolving model. We validate the effectiveness of our model by performing a comparative analysis with similar existing works. The advantages of our LFSPN model are twofold. Firstly, it offers a representation with two levels of abstraction: a graph representing the classic components of an SPN, and an additional layer encompassing the learning and migration aspects inherent to a mobile learning agent. Secondly, our model stands out for its flexibility and simulation simplicity.

Differential Evolution (DE) is a global optimization process that uses population search to find the best solution. It offers characteristics such as fast convergence time, simple and understood algorithm, few parameters, and good stability. To improve its presentation, we propose a differential evolution algorithm based on subpopulation adaptive scale and multi-adjustment strategy (ASMSDE). The algorithm separates the population into three sub-populations based on fitness scores, and different operating tactics are used depending on their characteristics. The superior population uses Gaussian disturbance, while the inferior population uses Levy flights. The intermediate population is responsible for maintaining the population's overall variety. The sizes of the three sub-populations are adaptively changed in response to evolutionary results to account for changes in individual differences over time. With the number of iterations increases and the disparities between individuals reduce, adopt a single population model instead of multi-population model in the later stage of evolution. The ASMSDE algorithm's performance is evaluated by comparing it to other sophisticated algorithms that use benchmark function optimizations. Experimental results show that the ASMSDE algorithm outperforms the comparison algorithms in the majority of circumstances, demonstrating its effectiveness and capacity to manage local optimum situations.

Uncertainty in robot dynamic systems is caused by model errors in the dynamic parameters, and accurate identification of the dynamic parameters is essential to improve the control accuracy of the robot. In this paper, a hybrid optimization strategy for modular robot manipulator dynamic model parameter identification is proposed to accurately identify the dynamic parameters of the robot manipulator. Firstly, the robot dynamics model with Coulomb viscous friction is established. Secondly, the cosine adaptive learning and reversal strategies are introduced to improve the genetic algorithm, and the improved genetic optimization algorithm is applied to optimize the excitation trajectories, and all the robot arm joints are commanded to follow the optimized excitation trajectories. In addition, considering that the Coulomb viscous friction model is not sufficient to accurately express the friction terms, a two-step identification method is proposed by analyzing the sensitivity of the parameters of the Stribeck friction model, combining the significantly identified friction coefficients with the quadratically optimized coefficients of the adaptive inverse genetic algorithm, which solves the problem of lower accuracy caused by the inaccuracy of the friction parameter identification. Then, the dynamic parameters are calculated using the least squares method to determine the system dynamics model information. Finally, the parameter identification and load identification are verified using a 6-degree-of-freedom modular robot manipulator, and the proposed hybrid optimization strategy effectively solves the defect of the low accuracy of the robot manipulator dynamics model compared to the dynamics model moment with Coulomb viscous friction, which in turn improves the control accuracy. Meanwhile, the load identification accuracy can reach 97% depending on the identified dynamics information.

Complex Fermatean fuzzy set (CFF-Sets) is one of the successful extensions of complex Pythagorean fuzzy sets (CPF-Sets). The main objective of the paper is to present complex Fermatean fuzzy sets (CFF-Sets), complex Fermatean fuzzy numbers (CFFNs) and some of their basic operational laws and their corresponding aggregation operators, which can represent the time-periodic problems and two-dimensional information in a single set. We introduce various novel operators, such as complex Fermatean fuzzy Einstein weighted geometric aggregation (CFFEWGA) operator, complex Fermatean fuzzy Einstein ordered weighted geometric aggregation (CFFEOWGA) operator, complex Fermatean fuzzy Einstein hybrid geometric aggregation (CFFEHGA) operator, induced complex Fermatean fuzzy Einstein ordered weighted geometric aggregation (I-CFFEOWGA) operator, and induced complex Fermatean fuzzy Einstein hybrid geometric aggregation (I-CFFEHGA) operator along with their structure properties, such as idempotency, boundedness and monotonicity. An illustrative example related to the selection of the more suitable location for hospital is to be considered to show the effectiveness and efficiency of the novel approach.

Wireless sensor networks (WSNs) are crucial in collecting environmental information through sensor nodes. However, limited energy resources pose a challenge, necessitating efficient routing algorithms to minimize energy consumption. Failure to address issues can consume energy and reduce network lifespan and overall efficiency. This research paper presents a cutting-edge approach for minimizing the consumption of energy within WSN through the implementation of an optimal routing method. The approach involves two steps: first, clustering sensor nodes using the pond skater algorithm (PSA) to select cluster head (CHs) for routing; second, by leveraging the ant colony optimization (ACO) algorithm, this study introduces an innovative technique that empowers a mobile sink to gather packets from given CHs and transmit effectively, send them back to the base station (BS). Notably, the authors make a significant contribution by introducing a different variant of the PSA algorithm to select CH. This novel approach aims to curtail the consumption of energy within WSN significantly. The authors also present an ACO-based head traversal for cluster method, resembling the traveling salesman problem coding, for minimized energy consumption. The study’s primary objectives include reducing energy consumption, minimizing packet delivery ratio, and prolonging the lifetime of the WSN. The assessment efficacy of the proposed method was achieved by regressive simulations using MATLAB on diverse scenarios. Through meticulous comparative analyses with several efficient algorithms, the method proposed here has shown significant performance in network lifetime comparison of PSACO in terms of Alive nodes with number of rounds PSO: 17.65%, GWO: 25%, CS: 33.33%, CBR-ICWSN: 66.66%, CCP-IC: 17.65%.

As a new metaheuristic algorithm, the Rat Swarm Optimization (RSO) has been increasingly applied to solve practical problems. However, RSO still suffers from slow convergence speed and easy trapping into local optima, especially for large-scale optimization problems. To overcome these drawbacks, a multi-strategy improved Rat Swarm Optimization algorithm with Whale Optimization Algorithm (MSRSO-WOA) is proposed. First, a segmented chaotic mapping is used to initialize the population to improve the quality of initial solutions. Second, a cosine oscillation weight is added to the position update process of the rat swarm, and new nonlinear exploration parameters and Levy flight development parameters are used to increase the convergence speed and exploration ability of the algorithm. Finally, the whale bubble spiral position update method of the Whale Optimization Algorithm is incorporated into RSO to improve the local search capability of the algorithm. The performance of MSRSO-WOA is evaluated by 23 well-known benchmark functions, 10 CEC testing functions, and 3 practical engineering problems. The results show that MSRSO-WOA has better optimization performance and stronger robustness than other compared algorithms.

Uncertain fractional differential equations fit more with the actual financial market since they have the non-locality features to mirror the memory and hereditary characteristics of the underlying asset price. In this paper, we investigate the option price in the asset price and volatility following the uncertain fractional differential equations in the sense of Caputo. Firstly, we propose the stock model with an uncertain fractional volatility and present the (alpha )-path of the uncertain fractional volatility model. Secondly, the pricing formulas of European and American options are obtained for the proposed model. Lastly, numerical experiments on market data are presented. Numerical calculations and data examples show the accuracy and efficiency of the proposed model.

Complementors who provide content on platforms are increasingly threatened by the entry of platform owners. Platform owners may enter the content market through offering vertically differentiated content either by self producing or hiring the complementor to produce. We build a game-theoretic model to analyze the platform owner’s entry decisions and the complementor’s response strategy considering the effects of demand complementarity, vertical content differentiation and consumer heterogeneity to both players’ strategies. We find that vertical content differentiation relaxes boundary conditions of entry, and it is more obvious when the platform owner has advantage in content value. However, we show that though the complementor may hold advantages on content value, price, or sales volume, it faces dependent dilemma once entry happens. Further, we demonstrate that second-party cooperation may mitigate the dependent dilemma and create a “win–win” situation through leveraging the platform owner’s efficiency in marketing and the complementor’s efficiency in content producing.