Advancements in Mathematical Methods for Pattern Recognition and its ApplicationsView this Special Issue
Research Article | Open Access
Xin Guan, Haiqiao Liu, Xiao Yi, Jing Zhao, "The Improved Combination Rule of D Numbers and Its Application in Radiation Source Identification", Mathematical Problems in Engineering, vol. 2018, Article ID 6025680, 10 pages, 2018. https://doi.org/10.1155/2018/6025680
The Improved Combination Rule of D Numbers and Its Application in Radiation Source Identification
The D numbers theory is a novel theory to express uncertain information. It successfully overcomes some shortcomings of Dempster-Shafer theory, such as the conditions of exclusiveness hypothesis and completeness constraint. However, the combination rule of D numbers does not satisfy the associative property, which leads to limitations in practical application for D numbers. In this paper, the improved D numbers theory is proposed to overcome the weakness based on the analysis of D numbers’ combination rule. A new algorithm is constructed with the strict proof to simplify the combination rule. The similarities and differences among DS theory, D numbers, and the improved D numbers are introduced with the numerical analysis. An illustrative example of the radiation source identification is presented to demonstrate the effectiveness of the improved method.
In real applications, uncertainty reasoning is widely applied to information fusion, risk assessment, pattern recognition, artificial intelligence, decision-making, and so forth. Many methods such as fuzzy sets theory, Dempster-Shafer theory, and possibility theory, have been proposed to model uncertain information.
Dempster  proposed the upper and lower probabilities that are induced by a multivalued mapping in the 1960s. His student Shafer introduced the concept of belief function and published the book , which is the sign of the formation of evidence theory. During the application process of evidence theory, many related problems have been studied in depth, such as the acquisition of basic probability assignments [3–6], information fusion of reliable sources [7, 8], conflict representation model [9, 10], and decision rule [11, 12]. Compared with the traditional Bayesian theory, DS theory, also called an evidence theory, needs weaker conditions and handles uncertain and incomplete information effectively, so it is often regarded as an extension of Bayesian theory. However, DS theory still has some shortcomings. When evidences acquired from different sources are highly conflicting, DS combination rule is defective in resolving evidence combination problem . In addition, there are some strong hypotheses on the frame of discernment. For example, the elements in the frame of discernment require mutual exclusiveness and the frame must be complete. To handle the existing shortcomings in the DS theory, a new theory is proposed by Deng , which is called D numbers.
The DS evidence theory combination rule cannot effectively deal with high-conflict information [9, 15]. A typical example is the zero trust paradox case proposed by Zadeh ; that is, 0 has one vote veto. For this problem, a large number of improved algorithms [16–21] made by relevant scholars are mainly divided into two categories: one is the improvement of combination rules and the other is the improvement of data sources [22–32]. The improvement of the combination rules is mainly divided into three categories. One is to recognize the multiplicative principle of the DS evidence theory and to study how to allocate the amount of conflict, such as Smets [25, 26], Yager [27, 28], and PCR1-6 [29–31]. The second type is the multiplicative rule of DS evidence theory, which gives additive combination rules, such as Murphy combination rules . The third category is to change the recognition framework and extend to the generalized power set. On this basis, new combination rules are given, such as the DSmT combination rule , which extends the recognition framework to the power set framework. The D number theory is a generalization of evidence theory. It cancels the assumption that the propositions in the recognition framework are mutually independent. Related scholars have done lots of research on it. Li  proposed a novel distance function of D numbers. Deng  introduced the difference between D numbers and DS theory based on generalized evidence theory.
On the application side, Zuo  analyzed investment decision, Deng  analyzed bridge condition assessment, and Deng  analyzed environmental impact assessment. Meanwhile, some methods were modified or extended by D numbers and have been applied to engineering field. Fei and Bian [38, 39] analyzed human resources selection and failure mode based on D numbers and TOPSIS. Su, Deng, and Fan [40–42] proposed the dependence assessment in human reliability analysis, supplier selection, and curtain grouting efficiency assessment based on D numbers and AHP. Deng and Zhou [43, 44] proposed the D-CFPR and D-DEMATEL theory based on D numbers. Liao  introduced transformer condition assessment using game theory and modified evidence combination extended by D numbers. Wang  introduced a multicriteria decision-making method based on fuzzy entropy and evidential reasoning with linguistic D numbers.
Though D numbers theory overcomes some defects of DS theory and is widely used in different fields, the combination rule of D numbers does not satisfy the associative property. In order that multiple D numbers can be combined correctly and efficiently, a combination operation for multiple D numbers is developed in . But it does not change the fact that D numbers’ combination rule does not preserve the associative property. In this paper, the improved D numbers theory is proposed based on the analysis of D numbers’ combination rule which satisfies the associative property. An illustrative example about radiation source identification is given to show the effectiveness of the proposed method.
The remainder of this paper is organized as follows. A brief introduction about the DS theory and D numbers is given in Section 2. Then, the concept of improved D numbers is depicted in Section 3. In Section 4, the similarities and differences among DS theory, D numbers, and the improved D numbers are represented. After that, an illustrative example is given to show the effectiveness of the proposed method in Section 5. Finally, conclusions are given in Section 6.
2. Materials and Methods Preliminaries
2.1. DS Theory
Let be a set of exhaustive and exclusive hypotheses, satisfying,The power set of is indicated by , and each element of is regarded as a proposition. It can be defined asBased on the two conceptions, mass function is defined as below.
Definition 1. A mass function is a mapping from to , formally defined bywhich should satisfy the condition:A mass function is also called a basic probability assignment (BPA), which measures the support degree of the proposition .
In real applications, for the same problem, there may be many different sources that acquire various evidences.
Definition 2. Considering two pieces of evidence from different and independent information sources, denoted by two BPAs and , the combination rule of DS theory is used to derive a new BPA from two BPAs, which is represented by , and defined as follows:withwhere is the conflict coefficient of two BPAs. Note that the combination rule of DS theory is only applicable to such two BPAs which satisfy the condition .
Definition 2′. Multiple pieces of evidence from different information sources also can be combined with the following formula. The result is a new piece of evidence, which incorporates the joint information acquired from various sources.withThe factor indicates the amount of evidential conflict. If , it shows that the sources are completely compatible. If , it shows that the sources are completely contradictory, and it is no longer possible to combine them.
2.2. D Numbers
D numbers theory is more suitable to the framework and is defined as follows.
Definition 3. Let be a finite nonempty set, and number is a mapping , withThe structure of the expression seems to be similar to mass function. However, the elements of do not require to be mutually exclusive, and D numbers theory is suitable to incomplete information.
Let , and a special form of D numbers can be defined as follows.or simply denoted aswhere and
Like the mass function, D numbers theory also has the combination rule to combine two D numbers.
Definition 4. Let be two D numbers, indicated byThe combination of , indicated by , is defined bywithwhereThe combination operation does not preserve the associative property. Based on formula (15), it is clear that, for ,while, for , As a result, a combination operation for multiple D numbers should be developed. Paper  represents a multiple D numbers’ rule of combination.
Definition 5. Let be numbers, is an order variable for each D number , indicated by tuple , and then the combination operation of multiple numbers is a mapping where is of the tuple which have the lowest .
In the meanwhile, for the special D number, an aggregation operator is defined as follows.
Definition 6. Let = , be a D number, and the integration representation of the D number is defined as
3. An Improved Combination Rule of D Numbers
As a novel theory to express uncertain information, D numbers’ combination rule does not preserve the associative property. For n D numbers, there are outcomes. Formulas (17) and (18) can be extended to multiple D numbers, and the outcome can be indicated bywhere is the parameter of the D numbers. From formula (21), it can be seen that the weight of the first combination number b is lower, which is . And the weight of the latter combination number b is higher, which is . The analysis shows that the weight is normalized, namely,The main reason for the imbalance of weight is that formula (15) involves addition and multiplication, which do not satisfy the associative property. In the meanwhile, the combination rule uses serial method to combine multiple numbers, as shown in Figure 1.
3.1. An Improved Algorithm for D Number Theory
In order for D numbers’ combination rule to satisfy the associative property, in this paper, the improved D numbers’ combination rule with parallel structure is proposed.
Before that, a novel algorithm is proposed to facilitate the elaboration of the problem. Letbe two summation formulas. With the new operation , the combination of can be indicated byThen, the following valid theorem can be acquired.
Theorem 7. The parameter of formula (15) can be rewritten aswhere
Theorem 8. The parameter of formula (25) can be further simplified, indicated bywhere
Definition 9. The improved D numbers’ combination rule is defined as follows.
Letbe D numbers. The combination of D numbers, indicated by , is defined bywithwhere
Theorem 10. Similarly, the parameter of formula (40) can be further simplified, indicated bywherewhen , then formula (41) can be reduced to formula (31). The proof of Theorem 10 can be carried out in a similar manner.
3.2. Evidence Combination of Time Series Based on Improved D Number Theory
In the case of obtaining n D numbers at the same time, the proposed method is able to satisfy the associative law. Unfortunately, the more common situation in practical engineering applications is that we first get a part of D numbers and then obtain the others and calculate their fusion results step by step. The following is a study of the D number combination rule for time series evidence.
In order to facilitate the explanation of the problem, it is worthwhile to record the parameter , , in formula (40) in which the m pieces of evidence are combined by the D number. Then, formula (40) can be simplified asSo, the problem of studying the D number combination rule of the time series evidence is converted to the known D number combination rule of the first n-1 moments,and the evidence at the nth moment,Solving the result of the D number combination rule at the first n time,It can be obtained by formula (40) thatSimilarly,For the parameter , it can be obtained by formula (24) that the multiplication-sum operation satisfies the commutative law. And so,By formulas (47)-(49), a general flow chart of the combination rule of the D number of time series evidence is given, as shown in Figure 2.
4. The Relationship among DS Theory, D Numbers, and Improved D Numbers
4.1. The Relationship between DS Theory and D Numbers
Though the D numbers derive from DS theory, the two theories solve different problems. The specific similarities and differences of the two theories are summarized as shown in Table 1.
Example 11. To make a comprehensive assessment of a student, we can get the following basic probability assignment based on DS theory.where is , is , is , and is the frame of discernment. The elements are mutually exclusive.
If we apply the D numbers theory to analyze the problem, we can get the expression as follows.where is , is , is , and is the frame of discernment. It is clear that the elements are not mutually exclusive.
4.2. The Contrast between D Numbers and Improved D Numbers
As we can see from formulas (17) and (18), the D numbers’ combination rule does not satisfy the associative property. However, for the improved D numbers, and have the same expression.namely,From formula (21), it can be seen that the weight of the first combination number b is lower, which is , and the weight of the latter combination number b is higher, which is . However, for the improved D numbers, the weights of the first combination number and the latter combination number b are the same, which is .
Example 12. Letbe 3 D numbers.
Then, we use the combination rules of D numbers and the improved D numbers separately to calculate the combined D number’s integration representation . And the results are shown in Table 2.
It can be seen from Table 2 that the algorithm can satisfy the commutative law, and the original D number theory does not satisfy the commutative law.
5. An Application in the Radiation Source Identification
In this section, in order to illustrate the application of the improved D numbers, a simple example about radiation source identification is given.
Radiation source identification mainly includes the analysis and decision-making of the detected signals. The detected signals of radiation source are radio frequency information, which have added up the noise and been modulated by the atmosphere. As a result, the radiation source identification can be regarded as an uncertainty reasoning process.
There are many theories to express uncertain information. In this paper, we will use the improved theory to identify the source of radiation. The problem is described in detail below.
There are 3 sources of information to investigate the target separately, indicated as . The 3 sources give corresponding decision based on different attributes. Among them, the measurable attributes of the source have distance , direction , and speed . The measurable attributes of the source have carrier frequency , pulse width , and pulse repetition period . The measurable attributes of the source have data chain , the frequency of communication , and modulation . Those three sources make enemy plane , friend plane , passenger plane , and unknown plane decisions. The decision results are shown as Table 3.
The accuracy and detection range of the source can vary with the battlefield environment and climate conditions. Table 4 shows the weights of sources in a specific time period.
In this paper, we use the improved D numbers to solve this problem. The application process is presented as follows.
Step 1. Express these assessment results of Table 3 in the forms of D numbers.
Take alternative as an example. For , in the opinion of source , the assessment results are indicated as , and the weights of attributes are indicated as . Then apply formula (11) to express these assessment results in the forms of D numbers. The expression is denoted byOther sources’ expressions can be required similarly, listed in Table 5.
Step 3. Apply formula (20) to calculate the integration representation of .
Step 4. Calculate and rank the integration representations of similarly. The ranking results are listed in Table 7.
Step 5. Analyze the data of Table 5 and make decisions. According to Table 7, the ranking of alternative is obtained, which is , where represents “better than”. It is clear that the target is the enemy plane .
From Table 7 it can be seen that the target of the original D number theory has similar results. As can be seen from the raw data, goal is more likely to be the final decision result. The improved algorithm proposed in this paper can effectively identify the target.
In this paper, the improved D numbers theory is proposed for radiation source identification. In the improved theory, a novel combination rule of D numbers theory is represented to satisfy the associative property. Meanwhile, a new algorithm is constructed with the strict proof to simplify the combination rule. For the application of radiation source identification, the novel theory is simpler and more effective. An illustrative example has shown the improved theory’s effectiveness.
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare no conflicts of interest.
Xin Guan and Haiqiao Liu conceived the concept and performed the research. Haiqiao Liu conducted the experiments to evaluate the performance of the proposed information fusion algorithm based on the improved D numbers method. Xiao Yi and Jing Zhao reviewed the manuscript. All authors have read and approved the final manuscript.
The work is partially supported by the special fund for Taishan scholar project: 201712072, National Natural Science Foundation of China: 61671463, National Natural Science Foundation of China: 91538201, Excellent Youth Scholar of the National Defense Science and Technology Foundation of China. Professor Xiao Yi provided many suggestions about the arithmetic operations. The Ph.D. students of the authors Jing Zhao and Haiqiao Liu in Naval Aviation University gave some valuable discussion about the definitions.
- A. P. Dempster, “Upper and lower probabilities induced by a multivalued mapping,” Annals of Mathematical Statistics, vol. 38, pp. 325–339, 1967.
- G. Shafer, A Mathematical Theory of Evidence, Princeton University Press, Princeton, NJ, USA, 1976.
- F. Valente and H. Hermansky, “Combination of acoustic classifiers based on dempster-shafer theory of evidence,” in Proceedings of the 2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07, pp. IV1129–IV1132, April 2007.
- Y. M. Zhu, L. Bentabet, O. Dupuis, V. Kaftandjian, D. Babot, and M. Rombaut, “Automatic determination of mass functions in Dempster-Shafer theory using fuzzy c-means and spatial neighborhood information for image segmentation,” Optical Engineering, vol. 41, no. 4, pp. 760–770, 2002.
- F. Salzenstein and A.-O. Boudraa, “Iterative estimation of Dempster-Shafer's basic probability assignment: Application to multisensor image segmentation,” Optical Engineering, vol. 43, no. 6, pp. 1293–1299, 2004.
- Y. Deng, W. Jiang, X. B. Xu, Q. Li, and D. Wang, “Determinging BPA under uncertainty environments and its application in data fusion,” Journal of Electronics (China), vol. 26, no. 1, pp. 13–17, 2009.
- T. Denoux, “Conjunctive and disjunctive combination of belief functions induced by nondistinct bodies of evidence,” Artificial Intelligence, vol. 172, no. 2-3, pp. 234–264, 2008.
- C. Fu and S. Yang, “The combination of dependence-based interval-valued evidential reasoning approach with balanced scorecard for performance assessment,” Expert Systems with Applications, vol. 39, no. 3, pp. 3717–3730, 2012.
- L. Hu, X. Guan, Y. Deng, D. Han, and Y. He, “Measuring conflict functions in generalized power space,” Chinese Journal of Aeronautics, vol. 24, no. 1, pp. 65–73, 2011.
- Y. He, L. F. Hu, X. Guan et al., “New conflict representation model in the generalized power space,” Journal of Systems Engineering and Electronics, vol. 23, no. 1, pp. 1–9, 2012.
- L. F. Hu, Y. He, X. Guan, D. Q. Han, and Y. Deng, “New probabilistic transformation of imprecise belief structure,” Journal of Systems Engineering and Electronics, vol. 22, no. 5, pp. 721–729, 2011.
- L. Hu, Y. He, X. Guan, Y. Deng, and D. Han, “A new probabilistic transformation in generalized power space,” Chinese Journal of Aeronautics, vol. 24, no. 4, pp. 449–460, 2011.
- L. A. Zadeh, “A simple view of the Dempster-Shafer theory of evidence and its implication for the rule of combination,” AI Magazine, vol. 7, no. 2, pp. 85–90, 1986.
- Y. Deng, “D numbers: theory and applications,” Journal of Information and Computational Science, vol. 9, no. 9, pp. 2421–2428, 2012.
- Y. He, L. F. Hu, X. Guan et al., “New conflict representation model in the generalized power space,” Journal of Systems Engineering and Electronics, vol. 23, pp. 1–9, 2012.
- C. Yu, J. Yang, D. Yang, X. Ma, and H. Min, “An improved conflicting evidence combination approach based on a new supporting probability distance,” Expert Systems with Applications, vol. 42, no. 12, pp. 5139–5149, 2015.
- W. Liu, “Analyzing the degree of conflict among belief functions,” Artificial Intelligence, vol. 170, no. 11, pp. 909–924, 2006.
- Z.-G. Liu, J. Dezert, Q. Pan, and G. Mercier, “Combination of sources of evidence with different discounting factors based on a new dissimilarity measure,” Decision Support Systems, vol. 52, no. 1, pp. 133–141, 2011.
- G. Sun, X. Guan, X. Yi, and J. Zhao, “Conflict Evidence Measurement Based on the Weighted Separate Union Kernel Correlation Coefficient,” IEEE Access, vol. 6, pp. 30458–30472, 2018.
- Y. Yang, D. Han, and C. Han, “Discounted combination of unreliable evidence using degree of disagreement,” International Journal of Approximate Reasoning, vol. 54, no. 8, pp. 1197–1216, 2013.
- A.-L. Jousselme, E. Bosse, and D. Grenier, “Measuring ambiguity in the evidence theory,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 36, no. 5, pp. 890–903, 2006.
- D. Dubois and H. Prade, “A set-theoretic view of belief functions,” Classic Works of the Dempster-Shafer Theory of Belief Functions, pp. 375–410, 2008.
- D. Dubois and H. Prade, “Representation and combination of uncertainty with belief functions and possibility measures,” Computational Intelligence, vol. 4, no. 3, pp. 244–264, 1988.
- G. Sun, X. Guan, X. Yi, and J. Zhao, “Belief intervals aggregation,” International Journal of Intelligent Systems, 2018, http://doi.org/10.1002./int.22046.
- P. Smets and R. Kennes, “The transferable belief model,” Artificial Intelligence, vol. 66, no. 2, pp. 191–234, 1994.
- P. Smets, “Data fusion in the transferable belief model,” in Proceedings of the 3rd International Conference on Information Fusion, FUSION 2000, pp. PS21–PS33, July 2000.
- R. R. Yager, “On the relationship of methods of aggregating evidence in expert systems,” Cybernetics and Systems, vol. 16, no. 1, pp. 1–21, 1985.
- R. R. Yager, “On the Dempster-Shafer framework and new combination rules,” Information Sciences, vol. 41, no. 2, pp. 93–137, 1987.
- F. Smarandache and J. Dezert, “Proportional conflict redistribution rules for information fusion,” in Proceedings of the USENIX Technical Conference, 2004.
- F. Smarandache and J. Dezert, “Modified PCR rules of combination with degrees of intersections,” in Proceedings of the 18th International Conference on Information Fusion, Fusion 2015, pp. 2100–2107, July 2015.
- J. Dezert and F. Smarandache, Advances And Applications of DSmT for Information Fusion Collected Works, American Research Press, Rehoboth, NM, USA, 3rd edition, 2009.
- C. K. Murphy, “Combining belief functions when evidence conflicts,” Decision Support Systems, vol. 29, no. 1, pp. 1–9, 2000.
- M. Li, Y. Hu, Q. Zhang, and Y. Deng, “A novel distance function of D numbers and its application in product engineering,” Engineering Applications of Artificial Intelligence, vol. 47, pp. 61–67, 2016.
- Y. Deng, “Generalized evidence theory,” Applied Intelligence, 2015.
- Q. Zuo, X. Qin, Y. Tian, and D. Wei, “A Multi-Attribute Decision Making for Investment Decision Based on D Numbers Methods,” Advances in Pure Mathematics, vol. 06, no. 12, pp. 765–775, 2016.
- X. Y. Deng, Y. Hu, and Y. Deng, “Bridge Condition Assessment Using D Numbers,” The Scientific World Journal, vol. 2014, Article ID 358057, 11 pages, 2014.
- X. Deng, Y. Hu, Y. Deng, and S. Mahadevan, “Environmental impact assessment based on D numbers,” Expert Systems with Applications, vol. 41, no. 2, pp. 635–643, 2014.
- L. Fei, Y. Hu, F. Y. Xiao, L. Chen, and Y. Deng, “A modified topsis method based on D numbers and its applications in human resources selection,” Mathematical Problems in Engineering, vol. 2016, Article ID 6145196, 14 pages, 2016.
- T. Bian, H. Y. Zheng, L. K. Yin et al., “Failure mode and effects analysis based on D numbers and TOPSIS,” Knowledge-Based Systems, vol. 2016, pp. 1–39, 2016.
- X. Su, S. Mahadevan, P. Xu, and Y. Deng, “Dependence assessment in human reliability analysis using evidence theory and AHP,” Risk Analysis, vol. 35, no. 7, pp. 1296–1316, 2015.
- X. Deng, Y. Hu, Y. Deng, and S. Mahadevan, “Supplier selection using AHP methodology extended by D numbers,” Expert Systems with Applications, vol. 41, pp. 156–167, 2014.
- G. Fan, D. Zhong, F. Yan, and P. Yue, “A hybrid fuzzy evaluation method for curtain grouting efficiency assessment based on an AHP method extended by D numbers,” Expert Systems with Applications, vol. 44, pp. 289–303, 2016.
- X. Deng, X. Lu, F. T. Chan, R. Sadiq, S. Mahadevan, and Y. Deng, “D-CFPR: D numbers extended consistent fuzzy preference relations,” Knowledge-Based Systems, vol. 73, pp. 61–68, 2015.
- X. Zhou, Y. Shi, X. Deng, and Y. Deng, “D-DEMATEL: A new method to identify critical success factors in emergency management,” Safety Science, vol. 91, pp. 93–104, 2017.
- R. Liao, H. Zheng, S. Grzybowski, L. Yang, Y. Zhang, and Y. Liao, “An integrated decision-making model for condition assessment of power transformers using fuzzy approach and evidential reasoning,” IEEE Transactions on Power Delivery, vol. 26, no. 2, pp. 1111–1118, 2011.
- J.-Q. Wang and S.-K. Huang, “Multi-criteria decision-making method based on fuzzy entropy and evidential reasoning with linguistic D numbers,” Kongzhi yu Juece/Control and Decision, vol. 31, no. 4, pp. 673–677, 2016.
Copyright © 2018 Xin Guan et al. 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.