Wireless Communications and Mobile Computing

Wireless Communications and Mobile Computing / 2021 / Article
Special Issue

Security, Trust and Privacy in Internet of Things

View this Special Issue

Research Article | Open Access

Volume 2021 |Article ID 5499173 | https://doi.org/10.1155/2021/5499173

Xiaoya Deng, "Multiconstraint Fuzzy Prediction Analysis Improved the Algorithm in Internet of Things", Wireless Communications and Mobile Computing, vol. 2021, Article ID 5499173, 7 pages, 2021. https://doi.org/10.1155/2021/5499173

Multiconstraint Fuzzy Prediction Analysis Improved the Algorithm in Internet of Things

Academic Editor: Chien-Ming Chen
Received12 Apr 2021
Revised28 May 2021
Accepted01 Jul 2021
Published15 Jul 2021

Abstract

Multiconstraint prediction is a research hotspot and difficulty in the Internet of Things. Aiming at the main problems existing in the process of multiconstraint fuzzy prediction, this paper studies the related key technologies and methods and proposes an improved multiconstraint fuzzy prediction analysis algorithm in the application of Internet of Things. This paper introduces the multiconstraint attribute analysis, multiconstraint attribute normalization processing, multiconstraint attribute weight processing, multiconstraint attribute prediction analysis granularity setting, etc. and systematically calculates the ambiguity of multiconstraint attributes and realizes the multiconstraint fuzzy prediction analysis algorithm. Finally, combined with relevant cases, the algorithm is compared with the existing research results, which shows the effectiveness and feasibility of the algorithm in the Internet of Things.

1. Introduction

The Internet of Things (IOT) refers to the real-time collection of any object or process that needs monitoring, connection and interaction, and the collection of various required information such as sound, light, heat, electricity, mechanics, chemistry, biology, and location through various devices and technologies such as information sensors, radio frequency identification technology, global positioning system, infrared sensors, and laser scanners. Through all kinds of possible network access, the ubiquitous connection between things and people can be realized and the intelligent perception, identification and management of goods and processes can be realized. The Internet of Things is an information carrier based on the Internet and traditional telecommunication network. It enables all ordinary physical objects that can be independently addressed to form an interconnected network [1].

The Internet of Things (IOT) is the “Internet connected with everything.” It is an extension and expansion network based on the Internet. It combines various information sensing devices with the network to form a huge network, realizing the interconnection of people, machines, and things at any time and any place [2].

The Internet of Things is an important part of the new generation of information technology. The IT industry is also called pan Internet, which means that things are connected and everything is connected. Thus, “the Internet of Things is the Internet of Things.” This has two meanings: first, the core and foundation of the Internet of Things is still the Internet, which is an extended network based on the Internet, and second, its client extends to any item to exchange information and communicate with each other. Therefore, the definition of the Internet of Things is a kind of network that connects any article with the Internet according to the agreed protocol through information sensing devices such as RFID, infrared sensors, global positioning system, and laser scanner for information exchange and communication, so as to realize the intelligent identification, positioning, tracking, monitoring, and management of articles [3].

Constrained optimization, namely, constrained optimization problem, is a branch of optimization problem. It is to find a group of parameter values under a series of constraints, so that the target value of a function or a group of functions can reach the optimal value. The constraints can be either equality constraints or inequality constraints. The key to find this group of parameter values is to meet the constraints and target values to achieve the optimal value. The methods for solving constrained problems can be divided into traditional methods and evolutionary algorithms.

The basic idea of traditional methods, such as the Newton method and gradient method, is to transform a dynamic object to a static object and multiobjective into a single-object one, from point and face search. The traditional methods have the following problems: (1) the traditional gradient-based optimization methods (such as the feasible direction method and constrained variable scale method) usually seek a feasible and descending direction first, then search the linear in this direction, and repeat the steps above to get the optimal solution of the problem. However, the optimal solution is often local optimal. (2) For many practical constrained optimization problems, on the one hand, because the objective function is often complex, it is not only the problem with high dimension but also the existence of several small points in the optimization surface, which makes the traditional gradient-based algorithm difficult to work. On the other hand, the objective function is often discontinuous or differentiable in practical problems. Some problems have no analytic expression, so traditional algorithms find it difficult to solve these problems. And (3) because of the existence of constraints, the feasible search space of decision variables is irregular (such as nonconvex and unconnected), which increases the difficulty of searching the optimal solution, and sometimes, it is difficult to find the feasible solution.

The evolutionary algorithm is an intelligent global optimization method, which requires very low properties of the function itself, often only requires that the objective function value can be calculated, and does not require it to have continuity, differentiability, and other analytical properties. At the same time, it is based on the population evolution algorithm, so it can use the evolutionary algorithm to solve constrained optimization problems. The key of using the evolutionary algorithm to solve the constrained optimization problem is how to deal with constraints effectively, that is, how to balance the search between feasible and infeasible regions.

Common evolutionary algorithms for solving constrained optimization problems include the penalty function method, genetic algorithm, evolutionary strategy, evolutionary programming, and ant colony algorithm and particle swarm optimization.

In general, whether the evolutionary algorithm can converge to the global optimal solution is independent of the initial population, while the traditional optimization method depends on the initial solution. The evolutionary algorithm has the ability of global search, but many traditional optimization methods often fall into local optimum. The evolutionary algorithm has a wide range of applications and can effectively solve different types of problems, while traditional optimization methods can only solve a certain type of problems.

The shortcomings are as follows: (1) the parameters of the evolutionary algorithm, such as population size, evolutionary algebra, recombination probability, and mutation probability, often need to be set according to experience and are related to the problem to a certain extent and (2) the convergence problem of the evolutionary algorithm, the convergence judgment of the evolutionary algorithm in solving practical problems, is the lack of theoretical guidance.

Multiconstraint problem is a difficult problem in the process of system engineering analysis. In many engineering fields, it is necessary to optimize the design and analysis of multiconstraint problems, especially complex system engineering [1, 2]. What is more complicated is that the development of complex systems is largely unknown due to the existence of multiple constraints. Therefore, it is very difficult to predict the development of complex systems scientifically, accurately, and reliably [3, 4], especially when the constraint information is fuzzy and uncertain. How to accurately capture information and make effective prediction, so as to provide good support for the design, maintenance, and repair service of the complex system, will have very important engineering application significance. At present, the research of fuzzy system prediction has analyzed some existing research results, especially literature [5, 6], based on the Grey Theory and neural network [79], vector machine, extension theory of genetic algorithm [1012], etc., which has good engineering application effect.

Due to the uncertainty of fuzzy information, how to measure fuzzy information effectively and make more detailed and accurate information prediction still need further research. Therefore, based on the fuzzy system theory [13, 14], this paper analyzes the multiconstraint fuzzy prediction problem and proposes an improved multiconstraint fuzzy prediction analysis algorithm and model [1519].

2. The Basic Theory

2.1. Multiconstraint Fuzzy Prediction Analysis Attribute Set

The content of multiple constraints has many meanings. First of all, in the process of design, analysis, and optimization of the complex system, the design objective is restricted by many factors. In order to get the ideal design result, we must deal with the relationship between design constraints and design objectives. Second, for complex systems, there are contradictions or incompatibilities among various types of design behaviors, design requirements, or design objectives. To deal with these contradictions or incompatibility, we must deal with the relationship between the factors involved. The third is the correlation between design attributes of complex systems. This association can be implicit or explicit, but it has an impact on the performance of complex systems. Therefore, in order to deal with the design problems better, we can extract the features and attributes of multiconstraints and form a multiconstraint fuzzy prediction analysis set, so that it can better predict and analyze the design objectives. Assuming that the extracted multiconstraint attributes are marked with the number , then, the multiconstraint fuzzy prediction analysis set is expressed as follows:

2.2. The Fuzzy Distance

In order to describe the proximity between the predicted analysis target and the fuzzy set , the fuzzy system theory gives the concept of the fuzzy distance. Suppose that the fuzzy value of the constraint attribute of the prediction analysis target corresponding to the fuzzy set is denoted as . The fuzzy interval of the fuzzy set about the constraint attribute is denoted as . Then, the fuzzy distance between the two is denoted as

In particular, for , it can be expressed as follows:

It is called the Hamming distance. For , it can be expressed as follows:

It is called the Euclidean distance.

3. Multiconstraint Fuzzy Prediction Analysis Algorithm and Model

3.1. The Normalization of Multiple-Constraint Attributes

Different predictive analytics indicators may have different dimensions. In order to make the results of the predictive analysis more reliable, all predictive and analytical indicators need to have a unified measurement standard. Since it is a fuzzy prediction analysis, for the generality of the discussion, it is assumed that the attribute values of the prediction analysis are all fuzzy intervals, that is, the initial magnitude of the prediction analysis target with respect to the prediction analysis index is and the normalized value is . The threshold value of the predictive analysis indicator is , so the normalized formula of the cost-type predictive analysis indicator can be expressed as follows:

The normalized formula of benefit prediction and analysis indicators can be expressed as follows:

3.2. Weight Handling of Multiple-Constraint Attributes

The multiconstraint attribute weights can be obtained by various methods, such as the AHP method, entropy weight method, comprehensive evaluation method, neural network method, and gray correlation method. AHP method has the advantages of simplicity, practicality, and reliable results. Therefore, this paper takes the AHP method as an example to illustrate the weight processing process of multiconstraint attributes. The process implementation is as follows [11]:(1)The first is forming a multiconstraint predictive analysis attribute set(2)Expert judgement is carried out using a 1–9 ratio scale to obtain a judgment matrix (3)Normalize the evaluation matrix (4)The maximum eigenvalue acquisition is performed on the evaluation matrix (5)Perform a consistency check on the evaluation matrix If , it indicates that the evaluation matrix has good consistency and the next step can be performed. Otherwise, the evaluation matrix needs to be reset(6)Weight acquisition

3.3. Multiconstraint Attribute Predictive Analysis Granularity Setting

The granularity setting of multiconstraint attribute predictive analysis is to set the interval division of the quantitative value of the multiconstraint fuzzy predictive analysis attribute set. The granularity should be selected appropriately, neither too large nor too small, and should be analyzed and selected according to the actual situation of the prediction analysis target. If the selected granularity is too small, the attribute set of multiconstraint fuzzy prediction analysis is divided too fine. In extreme cases, each data item is a reference target, which may lead to the failure to predict the target object. If the selected granularity is too large, the division of the multiconstraint fuzzy prediction analysis attribute set is too thick, which will lead to a too wide range of reference target and too wide range of prediction results, so that the target object cannot be accurately predicted.

3.4. Classical Domain Generation of Multiconstraint Properties

On the basis of setting the granularity of multiconstraint attribute prediction analysis, assuming that it is divided into classical domains, the quantitative value interval of the reference target of the corresponding classical domain on the attribute set of fuzzy prediction analysis can be obtained .

According to the magnitude of the interval , the reference objects corresponding to the historical records contained therein are determined. Assuming that the number of reference objects is , the classical field of the reference target formed for these reference objects is

In general, in order to ensure the reliability and accuracy of the prediction results, take the following values, even if the range is a little bit wider.

3.5. Multiconstraint Attribute Fuzzy Proximity Calculation

Assume that prediction target on the forecasting analysis index value for , according to the fuzzy distance calculation formula given above, the fuzzy distance between the prediction and analysis target , and the reference target can be obtained.

When the magnitude of the predictive analysis target with respect to the predictive analysis index is an exact value, that is, , it is described as follows:

The greater the fuzzy distance , the more distant the two are, and the smaller the fuzzy distance is, the closer they are. Therefore, the fuzzy proximity is introduced

Assuming that the weight of each predictive analysis attribute is taken into account, the comprehensive weighted fuzzy approximation is

3.6. Implementation of Multiconstrained Fuzzy Prediction Analysis

The larger the integrated weighted fuzzy proximity is, the closer the predicted analysis target is to the reference object, and vice versa, the farther away the predicted analysis target is from the reference object. Therefore, based on the comprehensive weighted fuzzy proximity , the predicted analysis target can be obtained according to the size of the integrated weighted fuzzy proximity and the corresponding domain value range of the corresponding reference target . Then, the fuzzy predictive value of the prediction analysis target must satisfy the following equation.

Thus, the realization process of the multiconstraint fuzzy prediction analysis algorithm can be shown in Figure 1.

4. Comparative Analysis

In order to verify the effectiveness and feasibility of the proposed multiconstraint fuzzy prediction analysis algorithm, the Euclidean distance calculation formula is used for analysis. At the same time, the implementation process of the algorithm and model proposed in this paper is explained.

Compared with the data in literature [5], the initial comparison data is shown in Table 1.


Classical fieldTarget range
Property 1Property 2Property 3

10–0.0700.135–0.1950.120–0.1800.070–0.150
20.070–0.1800.080–0.2700.110–0.2000.150–0.230
30.130–0.3400.200–0.2600.340–0.4000.230–0.300
Prediction object0.2470.2160.309/
Weight0.3000.4000.300

According to the calculation formulas of the fuzzy distance and fuzzy proximity, the corresponding calculation results can be obtained, as shown in Tables 2 and 3, respectively.


Classical field
Property 1Property 2Property 3

10.2150.0590.162
20.1340.1030.160
30.1060.0330.068


Classical fieldWeighted fuzzy approximation
Property 1Property 2Property 3

10.7850.9420.8380.864
20.8660.8970.8400.870
30.8940.9670.9320.937

It can be seen that the predicted result target interval of this algorithm should be 0.937. This is consistent with the results of literature [5]. However, it should be pointed out that the predicted object is closer to the target interval 1 for attribute 1 and closer to the target interval 2 for attribute 2. For attribute 3, the difference is not obvious, but because attribute 2 accounts for a large proportion, the difference should not be obvious, which is consistent with the calculation results.

Similarly, the same processing method was used for comparative analysis with literature [11] and the calculation results were shown in Tables 4, 5, and 6.


Classical fieldTarget range
Property 1Property 2Property 3Property 4

10.069–0.1580.085–0.3270.146–0.1960.077–0.1620.080–0.180
20.083–0.1810.172–0.2650.127–0.2750.106–0.2160.180–0.280
30.187–0.3270.256–0.2840.292–0.3320.209–0.3010.280–0.400
Object to be predicted0.1950.2020.2270.183/
The weight0.1540.4770.2880.081


Classical field
Property 1Property 2Property 3Property 4

10.0930.1210.0610.076
20.0800.0490.0780.059
30.0940.0690.0870.085


Classical fieldWeighted fuzzy approximation
Property 1Property 2Property 3Property 4

10.9070.8790.9390.9240.903
20.9200.9510.9220.9410.939
30.9060.9310.9130.9150.920

According to the fuzzy distance and fuzzy proximity calculation formula in the paper, corresponding calculation results can be obtained, as shown in Tables 2 and 3, respectively.

According to the data in the table, the target interval of the prediction results of this algorithm should be 0.939 corresponding to the target interval, that is, 0.180–0.280, which is consistent with the result of reference [11].

5. Conclusion

Due to the existence of fuzzy information, the prediction and analysis of complex systems under multiple constraints become very complex. Therefore, the study of multiconstraint fuzzy prediction analysis method has important engineering value. An improved multiconstraint fuzzy prediction analysis algorithm is proposed. The algorithm focuses on the establishment of the attribute constraint set, standardization, weight and prediction analysis of the granularity set, generation of the classical domain, calculation of fuzzy distance and fuzzy closeness, constraint fuzzy prediction analysis, and implementation of the algorithm and model. At the same time, in order to verify the effectiveness and feasibility of the proposed algorithm, this paper compares and analyzes the proposed algorithm and related algorithms. The results are consistent, which provides a new solution and method support for multiconstraint fuzzy prediction analysis.

Data Availability

The labeled dataset used to support the findings of this study is available from the corresponding author upon request.

Conflicts of Interest

The author declares no competing interests.

Acknowledgments

TangSeng Huang helped perform the analysis with constructive discussions. This study is supported by the Sichuan Provincial Department of Education project (18ZA0419), the educational reform project of Sichuan University of Arts and Science (2020JY014), the and first-class course project of Sichuan University of Arts and Science (2020KCC026).

References

  1. L. Zhang, W. Tang, Y. Liu, and T. Lv, “Multiobjective optimization and decision-making for DG planning considering benefits between distribution company and DGs owner,” International Journal of Electrical Power & Energy Systems, vol. 73, pp. 465–474, 2015. View at: Publisher Site | Google Scholar
  2. T. Dun, Y. Jun, and D. Min, “Object-oriented multiple poles and multi-objective constraint optimization design method,” Machine Building & Automation, vol. 4, pp. 1–4, 2016. View at: Google Scholar
  3. L. Peng, H. Min, and H. Ke, “Research on the human resource allocation of scientific research projects based on multi-objective optimization,” Computer Applications and Software, vol. 34, no. 5, pp. 217–222, 2017. View at: Google Scholar
  4. Z. Su, “Prediction of demands on cultivated land based on multi-criteria decision,” Hubei Agricultural Sciences, vol. 53, no. 17, pp. 4219–4243, 2014. View at: Google Scholar
  5. L. Ya, L. Jun, L. Jin, and X. Xiao, “Research and application of mobile phone gray fault rate based on gray model,” Computer Applications and Software, vol. 35, no. 2, pp. 112–116, 2018. View at: Google Scholar
  6. Z. Zhi, T. Nan, and Y. Hong, “On analysis and prediction of schooling population in compulsory education of Tianjin based on multi-factor gray model and recursive calculation method,” Journal of Southwest China Normal University (Natural Science Edition), vol. 42, no. 3, pp. 49–55, 2017. View at: Google Scholar
  7. L. Wei, G. Xiao, and M. Yin, “Research and application of uncertain genetic neural network in landslide hazard prediction,” Computer Engineering, vol. 43, no. 2, pp. 308–316, 2017. View at: Google Scholar
  8. X. Ting, W. Wen, and D. Liang, “Construction land demand forecasting of Chengdu city based on PCA-RBF neural network,” Journal of Southwest University (Natural Science), vol. 39, no. 11, pp. 183–190, 2014. View at: Google Scholar
  9. Z. Pan, X. Shou, and L. Jun, “A prediction model of water content of fresh tobacco leaves based on support vector machine regression,” Journal of Southwest University (Natural Science), vol. 38, no. 4, pp. 165–170, 2016. View at: Google Scholar
  10. Y. Jing, B. Qiang, L. Jie, and W. Fu, “Construction of aggregation quality predicting model for digital resource in library based on improved genetic algorithm and BP neural network,” Data Analysis and Knowledge Discovery, vol. 1, no. 12, pp. 49–62, 2017. View at: Google Scholar
  11. M. Zhou, Y. Long, W. Zhang et al., “Adaptive genetic algorithm-aided neural network with channel state information tensor decomposition for indoor localization,” IEEE Transactions on Evolutionary Computation, p. 1, 2021. View at: Publisher Site | Google Scholar
  12. Q. Guo and G. Zou, “Prediction methods for extension architecture programming based on decision tree classification,” CAAI Transactions on Intelligent Systems, vol. 12, no. 1, pp. 117–123, 2017. View at: Google Scholar
  13. S. Li and A. X. Ou, “The effectiveness students’ ideological evaluation model of university and political education based on the fuzzy mathematics theory,” Mathematics in Practice and Theory, vol. 45, no. 13, pp. 299–306, 2015. View at: Google Scholar
  14. M. M. Silva, A. P. de Gusmão, T. Poleto, L. C. Silva, and A. P. Costa, “A multidimensional approach to information security risk management using FMEA and fuzzy theory,” International Journal of Information Management, vol. 34, no. 6, pp. 733–740, 2014. View at: Publisher Site | Google Scholar
  15. M. Zhou, X. Li, Y. Wang, S. Li, Y. Ding, and W. Nie, “6G multi-source information fusion based indoor positioning via Gaussian kernel density estimation,” IEEE Internet of Things Journal, p. 1, 2020. View at: Publisher Site | Google Scholar
  16. M. Zhou, Y. Wang, Y. Liu, and Z. Tian, “An information-theoretic view of WLAN localization error bound in GPS-denied environment,” IEEE Transactions on Vehicular Technology, vol. 68, no. 4, pp. 4089–4093, 2019. View at: Publisher Site | Google Scholar
  17. M. Zhou, Y. Wang, Z. Tian, Y. Lian, Y. Wang, and B. Wang, “Calibrated data simplification for energy-efficient location sensing in internet of things,” IEEE Internet of Things Journal, vol. 6, no. 4, pp. 6125–6133, 2019. View at: Publisher Site | Google Scholar
  18. S. K. Shiroodi, M. Ghafoori, H. R. Ansari, G. Lashkaripour, and M. Ghanadian, “Shear wave prediction using committee fuzzy model constrained by lithofacies, Zagros basin, SW Iran,” Journal of African Earth Sciences, vol. 126, pp. 123–135, 2017. View at: Publisher Site | Google Scholar
  19. Y. Yan and B. Su, “Design of explicit fuzzy prediction controller for constrained nonlinear systems,” Mathematical Problems in Engineering, vol. 2015, Article ID 857546, 7 pages, 2015. View at: Publisher Site | Google Scholar

Copyright © 2021 Xiaoya Deng. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder
Views43
Downloads97
Citations

Related articles

We are experiencing issues with article search and journal table of contents. We are working on a fix as to remediate it and apologise for the inconvenience.

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.