نشریه سیستم های فازی و کاربردها
سال چهارم شماره 2 (پیاپی 9، پاییز و زمستان 1400)

In this regard, in the framework of deep learning and based on inaccurate information obtained from images of retinal layers, an end-to-end network has been trained to analyze and segment retinal layers. The result of this study is the introduction of a kit that can be installed on smartphones, which can be used to easily and simply monitor users' vision.The input of this structure is OCT images of the eye, without the need to extract hand-crafted features as input from the image, and its output is accurate spatial information of the retinal layers. The time required for each scan of OCT input images in this kit is about 10ms, which makes it possible to use such kits in a variety of smartphones easily. This kit, which assesses the health of ocular arteries, is inexpensive and can be used for remote locations or people who are not able to visit specialized eye centers or need frequent and continuous monitoring of their eye health. The ranking results presented in the present article show that the proposed structure has an acceptable performance and is very close to the average opinions of the group of specialist ophthalmologists. The results of evaluations and comparisons are made within the framework of multi-criteria decision-making approach. In order to extract the results in this framework, the TOPSIS algorithm is used in which the weights of the indicators are determined based on the entropy size.

Hypotheses testing is an effective technique for decision making on the manufacturing process capability. Considering fuzzy quality rather than precise specification limits, we can make more reliable decisions for investigating the manufacturing process capability. In this paper, an applied study on the basis of fuzzy quality using the Yongting’s index is presented. The proposed approach which is used in this applied study is a technique for testing the capability of a normal process in producing products within the preset fuzzy specification limits. The adversity to test process capability indices is complexity in the distribution of its natural estimator, even under the Normal distribution. Also, there is the challenge in testing process capability based on fuzzy quality. The non-parametric approach which is used to test the performance of a product based on fuzzy quality and random sampling techniques which is presented based on the Monte Carlo simulation method and can be generalized for various fuzzy qualities. This study is presented to investigate the quality of paint thickness in the automobile polishing process based on trapezoidal fuzzy quality. The numerical computations are presented to show the performance of the Monte Carlo simulation method for making reliable decisions in testing the Yongting’s index.

Project managers often face different and sometimes conflicting goals in optimizing project resources. In recent years, the demand of the project stakeholders to reduce the total cost of the project and reduce the time and increase its quality have intensified. Therefore, the balance of time, cost, and quality are the project's main goals. These topics lead researchers to develop models that add a quality factor to previous time balance models. This research provides a model for balancing cost, time and, quality, in discrete mode. Of course, since the quality factor is not a quantitative factor-like cost and time factors, and in the real world of projects, quality is a linguistic variable. It is obtained from the point of view of experts, so in this research, the quality of activities is shown as fuzzy numbers. By finding the optimal combination (time, cost, quality) of each activity, this model compresses the activities to reduce the costs of the whole project, while achieving the desired quality. A new metaheuristic algorithm called NHGA is introduced to solve the model, which has a much higher efficiency than GA to solve model, so with a case study, the efficiency of the proposed algorithm and the flexibility of the proposed model for project managers is shown

In this paper, a non-penalty differential inclusion-based model is proposed for solving nonsmooth convex optimization problems subject to linear equality and bound constraints. We prove the convergence of the trajectories to the equality feasible region in finite time. Also, the equivalence of the equilibrium point of the proposed neural network and the optimal solution of the original optimization problem is shown. In addition, we prove the stability of the proposed neural network in the sense of Lyapunov and the global convergence to an exact optimal solution to the original problem. In comparison with some existing models for solving nonsmooth convex optimization problems, there does not exist any penalty parameter or penalty function in the model's structure and the implementation of the proposed model is easier. In the end, as an application, the proposed neural network is reduced to a model for solving nonsmooth convex optimization problems subject to linear equality and $xgeq0$ constraints. Also, some illustrative examples are given to show the effectiveness of the proposed neural networks.

In this paper we investigate fuzzy tree automata. Fuzzy tree automata extend fuzzy automata. In fuzzy automata we deal with strings and in fuzzy tree automata we utilize trees instead of strings. We define a recognizable step fuzzy tree language and by a theorem we show that recognizable fuzzy tree language is recognizable step fuzzy tree language. Then, we define a deterministic fuzzy tree automaton and support of a fuzzy tree language and show that support of a fuzzy tree language can be recognized by a fuzzy tree automaton. We explain these concepts by examples. We also show that for a recognizable step mapping, there exists a deterministic fuzzy tree automaton that can recognize it, and vice versa. In fact we prove that the class of recognizable fuzzy tree languages is equal to the class of normal recognizable step mappings.The obtained theorems generalize the results of weighted automata to tree automata.

Permanent magnet synchronous machines (PMSM) used in spacecraft must be able to control the magnet flux. For this purpose, in addition to the magnet, an excitation coil is placed on their rotor, which is widely used in control purposes. Reducing torque ripple and increasing the average torque are two important goals in the design of electric machines. In this paper, first the average torque and torque ripple for the PMSM in several cases are determined by numerical method. Then, analytical functions are obtained for them by using multilayer perceptron neural network. Fixed coefficients in the neural network are adjusted so that not only the errors for the training, test, validation, and total data, are very small but also the error for the test data is very close to the error for the training data. After that, by multi-objective fuzzy particle swarm optimization algorithm, the dimensions of the PMSM are determined in such a way that the torque ripple is minimum and the average value of torque is maximum. The determination of the optimal point at the pareto front will be determined by using the fuzzy maximum-minimum method. In this paper, numerical analysis is performed with Maxwell software and neural network, optimization and fuzzy are modeled by Matlab one.

Organizations and institutions have internal departments that the overall performance of the organization is the result of the performance of each of these departments or stages. Each stage has its own input and output factors as well as connecting factors between stages. The factors between stages are called intermediate indicators. In a two-stage structure, the intermediate indicators are the outputs of the first stage, which are used as the input of the second stage. Evaluation performance of an organization should be determined by considering the performance of each of its stages. Data envelopment analysis is one of the appropriate methods to evaluate performance based on several indicators. In practice, it is not possible to determine these indicators with exact values. In this paper, we proposed a model for determining the performance of two-stager systems in a fuzzy environment, also as well as the allocation of resources to it. We apply inverse data envelopment analysis and propose a multi-objective programming model that, by increasing the output of the unit under evaluation, determines the rate of increase of the first stage inputs and intermediate indicators in a way that maintains its efficiency. Then we illustrate the proposed model to allocate resources to bank branches.

The purpose of this paper is to express and calculate the variance confidence interval with the Chachi method of data conversion and without converting the initial data with the bootstrap method. Because the data conversion depends on the amount of variance in the denominator of the conversion expression, it fluctuates and as a result, estimating the confidence intervals and testing the hypothesis is challenged. The results of both methods are compared and their advantages and disadvantages are examined. For this purpose, they first deal with concepts such as the fuzzy random variable and the distance between fuzzy data based on the concept of α-shak as well as α -cut. The hypothesis testing process is then performed for the variance of a sample fuzzy data. The bootstrap method was used to test the hypothesis and the mentioned confidence intervals were compared with the confidence intervals obtained by α -bresh method.

The purpose of this paper is to ‎appl‎y ‎of ‎concept ‎of‎ fuzzy ‎bridge ‎and ‎to ‎ex‎tend ‎of ‎notation ‎of‎ fuzzy domination ‎number‎ in fuzzy ‎graph. ‎‎‎F‎uzzy ‎bridge ‎introduces a‎ ‎new ‎notation ‎of‎ fuzzy ‎domination ‎sets ‎and‎ fuzzy domination ‎numbers ‎such ‎that ‎is ‎consedering‎‎ ‎for‎ fuzzy ‎graphs ‎with ‎cyclics ‎and ‎vertices ‎are ‎l‎ocated‎ ‎on ‎cyclics. ‎To check the ‎importance ‎of‎ fuzzy domination ‎numbers ‎based ‎on‎ fuzzy ‎bridge, ‎we ‎compare ‎it‎ ‎to ‎other ‎distinct‎ fuzzy domination ‎numbers ‎and ‎show ‎that‎ fuzzy domination ‎numbers ‎based ‎on‎ fuzzy ‎bridge ‎is ‎optimal.‎‎‎‎‎The main method‎, ‎in this research is based on computations of fuzzy domination number of complete fuzzy graphs with distinct fuzzy vertices and ‎‎‎is ‎worked ‎on ‎spec‎ial ‎and ‎useful‎ ‎fuzzy ‎graphs ‎such ‎as ‎st‎rong ‎2-‎wh‎eel ‎fuzzy‎ ‎graph ‎and ‎2-‎complet‎e ‎fuzzy ‎graph. ‎The ‎study ‎on‎ ‎2-‎wh‎eel ‎fuzzy‎ ‎graph ‎and ‎2-‎complet‎e ‎fuzzy ‎graph‎ was given for the first time in this ‎paper ‎and ‎are ‎worked ‎via ‎the ‎generalize ‎of ‎2-part ‎of‎ ‎wh‎eel ‎fuzzy‎ ‎graph ‎and ‎ ‎complet‎e ‎fuzzy ‎graph ‎which ‎these ‎special‎ fuzzy ‎graphs ‎have ‎many ‎applications ‎on ‎complex ‎networks.‎ ‎In ‎this ‎study,‎ computations of fuzzy domination number of ‎special‎ fuzzy graphs ‎are ‎done‎ ‎‎‎based ‎on ‎fuzzy ‎bridge ‎with ‎distinct ‎vertices.‎ ‎‎‎‎The paper includes implications for the development of fuzzy graph‎, ‎and for modeling the uncertainty problems ‎and ‎applicaions ‎in ‎real ‎world ‎which ‎we ‎consider ‎t‎wo‎ ‎modes‎ of real ‎problems.‎