Picture Fuzzy N-Soft Sets and Their Applications in Decision-Making Problems

In this article, firstly, we describe picture fuzzy N-soft sets (PFN-SSs) as a generalization of picture fuzzy sets (PFSs) and N-soft sets (N-SS) by observing that one of the essential concept of neutral grade is missing in intuitionistic fuzzy N-SS (IFN-SS) theory. The concept of neutrality grade can be observed in the situation when we encounter human views including more answers of type: yes, abstain, no, refusal. For instance, in election the election commission or election council issues voting papers for the candidate. The voting outcomes are categorized into 4 groups with the number of papers namely, vote for, abstain, vote against, and refusal voting. Further, We define the fundamental properties of PFN-SS and introduce M-subset, F-subset, compliment, intersections, unions, of PFN-SS and give their examples. Secondly, we define an algorithm to cope with PFN-SS data which is more generalized then the algorithm defined for IFN-SS. To show the advantage and usefulness of the defined technique, we give two examples from real life by utilizing PFN-SS data. The result shows in the comparison that our initiated method is more general and suitable than the IFN-SS, fuzzy N-SS (FN-SS), and N-SS.


Introduction
The majority of the real-life problems include uncertainty and vagueness in data. Handling such data is a big issue in many fields like engineering, social sciences, medical sciences, and economics. In light of this worry, Zadeh [1] presented fuzzy sets (FSs). The idea of FS is truly manageable because it catches vague and uncertain data. In FSs theory the grade of membership belongs to a closed interval [0,1]. Nonetheless, it may not generally be right to assume that the grade of non-membership in FS is equivalent to 1 minus the grade of membership, due to the presence of the degree of uncertainty or indeterminacy. Consequently, Atanassov [2] defined the idea of intuitionistic FSs (IFSs), which is a suitable generalisation of FSs. After this, Cuong [3] characterised a new idea by putting an additional grade which is a grade of neutral membership, and named it to picture FS (PFS). It triply contains positive membership, neutral membership, and negative membership. Researchers show interest to work on PFS because it is straightforwardly applied to tackle everyday life issues. Some fuzzy logic operators for PFSs were given by Cuong and Van Hai [4]. Luo and Zhang et al. [5] described new similarity measures for PFSs and its applications. Khan et al. [6] presented bi-parametric distance and similarity measures of PFSs and their applications in medical diagnosis. Khan et al. [7] presented distance and similarity measures for spherical FSs and their application in selecting mega projects. Mahmood et al. [8] gave the concept of complex hesitant FSs (CHFSs). The idea of complex dual hesitant FSs (CDHFSs) was given by Mahmood et al. [9].
Nonetheless, Molodtsov [10] showed that few challenges of the previously mentioned notions can be overwhelmed with the establishment of parameterised interpretation of the universe, an idea giving rise to soft sets (SSs) [10]. Soon, Maji et al. [11] established few algebraic operations in SS theory, which currently assume a significant part in fields such as DM [12][13][14][15][16] and data analysis [17]. Ali et al. [18] defined some modified operations in SS theory. Anyway, the information frequently comes across in a complex structure. A lot of researchers have developed new models to overcome such types of data and merge the advantages of the presented models. These models are developed to resolve a vast range of DM issues since they have more advantages than the original models. For example, by the mixture of FSs and SSs, Maji et al. [19] defined a new model, called fuzzy SS (FSS). A new idea of DM issues based on FSSs was given by Alcantud [20]. Maji et a. [21] gave the idea of intuitionistic FSSs (IFSSs) which is a mixture of IFSs and SSs. Khan et al. [22] defined a novel approach to generalised IFSSs and its application in decision support system. Khan et al. [23] gave another view on generalised interval-valued IFSSs. The renewable energy source selection by remoteness index-based VIKOR method for generalised IFSSs was presented by khan et al. [24]. The distance and similarity measures of generalised IFSSs were defined by khan and Kumam [25]. Yang et al. [26] defined an adjustable soft discernibility matrix based on PFSSs and its application in DM. Khan et al. [27] defined generalised PFSSs and described some modified operations for PFSSs. Jan et al. [28] defined multi-valued PFSSs and their applications in group DM problems. The applications of generalised PFSSs in concept selection were given by khan et al. [29]. Khan et al. [30] described an adjustable weighted soft discernibility matrix based on generalised PFSS and its applications in decision-making. Khan et al. [31] defined a theoretical-justifications for the empirically successful VIKOR approach to multi-criteria DM. Akram et al. [32] described hybrid DM frameworks under complex spherical FSSs.
We noted a lot of modifications of SSs in various directions such as probabilistic SS theory [33], bijective SSs [34], and fuzzy bipolar SSs [35]. Specifically, Fatimah et al. [36] introduced N-SS as a modification of SS that takes into consideration multinary parameterised interpretation of the universe. Alcantud et al. [37] interpreted an N-SS approach to rough sets. Kamaci and Petchimuthu [38] defined bipolar N-SS theory with applications. Fatimah and Alcantud [39] presented the multi-fuzzy N-SS and its applications to DM. Akram et al. [40] defined complex neutrosophic N-SS. A hybrid DM approach under complex Pythagorean fuzzy N-SS was given by Akram et al. [41]. Mahmood et al. [42] defined the complex picture fuzzy N-SSs. Akram et al. [43] provided the parameter reductions in N-SSs and their applications in DM. The inspiration for their defined N-SS is given as. An investigation of models with SSs as one of their components had indicated that the authors were motivated by SSs and their work of assessments (memberships are either 0 or 1) or the situation with partial memberships containing in [0,1] [17,44]. Anyway, a lot of real-life DM issues contain information with discrete and non-binary structure, which consequently didn't in the shape of those formats. It is simple to create examples including non-binary assessments. We frequently get them in ranking or rating position such as websites comparing software, movies, or mobile phones. Even though they frequently appear as several stars or dots (such as four stars, three stars, two stars, one star), which can simply be linked with natural numbers. As additional evidence of the capability of N-SS theory, Akram et al. [45] merge the N-SS with FS to defined FN-SS. After this, Akram et al. [46] defined the notion of IFN-SS which the mixture of IFS and N-SS.
IFN-SS is a good tool to deal with uncertainty and is in the form of (K, ((I, T, N))), where K is an intuitionistic fuzzy number and (I, T, N) is N-SS. It is noticed that one of the essential notion of neutrality grade is missing in IFN-SS theory. The notion of neutrality grade can be observed in the situation when we encounter human views including more answers of kinds: yes, abstain, no, and refusal. For instance, in an election, the election commission or election council issues ballot papers for the contestant. The voting outcomes are categorised into 4 groups with the number of papers, i.e. vote for, abstain, vote against, and refusal voting. Class of abstain signifies that the voting paper is a white paper refusing both agree and disagree for the contestant but still takes the vote. Class of refusal voting is either unacceptable papers or avoiding the vote. Nevertheless, the neutral grade can be taken into the consideration in medical diagnosis. For example, there may not influence the symptoms of fever, leg pain on the diseases heart problem, and brain tumour. Likewise, the symptoms of heart problems and blood cancer have a neutral influence on the diseases flu, corona, TB, etc. In IFN-SS we lost some of the important information in the shape of neutral grade which effect the end result in the decision-making problem. To overcome this and other weaknesses presented in some existing models in this article, we presented the notion of PFN-SS which is the extension of IFN-SS, FN-SS and will perform better than IFN-SS in DM. The geometrical representation of the PFN-SS is given in Figure 1.
The organisation of this article as follows: • Section 2 is consists with preliminaries in which we recall few basic definitions such as FSs, IFS, PFSs, SSs, FSS, IFSS, PFSS N-SSs, and IFN-SS. We also recall the basic properties of some of the above basic definitions. • Section 3 is a picture fuzzy N-Soft set in which we initiate PFN-SSs which is a mixture of PFSs and N-SS and explain it by giving examples. We also initiate the fundamental properties of PFN-SS in Section 3. • Section 4 of this article is applications of PFN-SS in which we introduce an algorithm to cope with PFN-SS data and to show the advantage and usefulness of the defined technique. • Section 5 is a comparison in which we do a comparison of our initiated model with IFN-SS. • In Section 6, the conclusion and future study of the proposed work are presented.

Preliminaries
In this section, let us recall few basic definitions such as FSs, IFS, PFSs, SSs, FSS, IFSS, PFSS N-SSs, and IFN-SS. We also recall the basic properties of some of the above basic definitions. U = ∅ be set of universe and A be set of attributes throughout this manuscript.
Zadeh [1], defined the idea of FS, which gives a powerful structure to dealing with uncertainty.
is a grade of membership. For every u ∈ U, ξ D (u) describes the degree to which u be a member of FS D. The pair (u, ξ D (u)) is called fuzzy number (FN) In [1] Zadeh described some basic operations which are as follows: u)) and E = (u, ξ E (u)) be two FNs. Then In [2], Atanassov defined the concept of IFS Definition 3: [2] An IFS over U is given as where ξ D : U → [0, 1] is a grade of membership and ϑ D : U → [0, 1] is the grade of non-membership with a condition that 0 ≤ ξ D (u) + ϑ D (u) ≤ 1 for each u ∈ U. The pair (u, (ξ D (u), ϑ D (u))) is called intuitionistic FN (IFN).
In [3], Cuong described PFSs by putting an additional grade of membership called the grade of neutral membership. Mainly, the notion of PFS may be appropriate in the environment when one faces human judgments containing more replies of the form: yes, no, abstain refusal. Voting is the best example of PFS. Definition 4: [3] A PFS over U is given as where ξ D : U → [0, 1], is a grade of membership, δ D : U → [0, 1] is the grade of neutral membership and ϑ D : U → [0, 1] is the grade of non-membership with a condition that In [3] Cuong also defined basic operations of PFS which are given below , ϑ E (u))) be two PFNs. Then In [10], Molodtsov described the SS which is entirely a new tool to cope with the uncertainties from a parametrisation opinion. Most of the time the parameter set contains characteristics, attributes, or properties of the alternatives. Maji et al. [19] described the FSS, which is a mixture of FS and SS. In FSS, each parameter should be represented by grade of membership, as in real-world, the opinion of the people is represented by some degree of imprecision. Maji et al. [21] described the IFSS, which is a mixture of IFS and SS. In [26], Yang described the PFSS, a mixture of PFS and SS. In PFSS, we can view vagueness from a parametrisation opinion in a picture fuzzy (PF) environment. Fatimah et al. [36] described the idea of N-SS which is the modification of SS.
Fatimah et al. [36] also described the extended and restricted intersection and union of N-SS which is given below Definition 11: [36] Let 1 = (I 1 , T 1 , N 1 ) and 2 = (I 2 , T 2 , N 2 ) be two N-SSs with T 1 ∩ T 2 = ∅. Then the restricted union and the intersection of 1 and 2 are described below: Definition 12: [36] Let 1 = (I 1 , T 1 , N 1 ) and 2 = (I 2 , T 2 , N 2 ) be two N-SSs. Then the extended union and the intersection of 1 and 2 are described below [46], defined IFN-SS.

Picture Fuzzy N-soft Sets
In this part of the article, we describe the notion of PFN-SS and present some basic operations on PFN-SS such as union, complement, intersection, and so on. PFN-SS is a hybrid model of PFS and N-SS. In this article, U represent the universe set, A be set of attributes, Y = {0, 1, 2, . . . , N − 1} be a set of ordered grades and = (I, T, N) be N-SS over U, where T ⊆ A. Now we describe the definition of a PFN-SS and give an example to illustrate it.  In the above definition, for each parameter (attribute) the mapping Y allot a PFS on the image of that attribute under the mapping I. Consequently, for every ∈ T and u ∈ U∃ a unique (u, } be a set of five corona vaccines which is under consideration of a health expert to buy and T = { 1 , 2 , 3 } be a set of attributes, where 1 represents price, 2 represents experimental results of vaccines, and 3 represents delivery time. Based on these attributes the health expert assigns grades to the corona vaccines in the form of stars which is given in Table 1. Where Four stars are for 'Excellent', three stars are for 'very good', two stars are for 'good', one star is for 'normal', a hole is for 'poor'. The set Y = {0, 1, 2, 3, 4} can be associated with the stars given in Table 1, where 4 represents , 3 represents , 2 represents , 1 represents , and 0 represents . The 5-SS (I, T, 5) is given as follows: The tabular form of 5-SS is interpreted in Table 2. The membership (ξ ), neutral (δ), and non-membership (ϑ) follow the following grad-  (I, T, 5) Table 3. The tabular representation of the picture fuzzy 5-soft set (Y, (I, T, 5)) in Example 1.
Clearly, the PF5-SS can be described in the tabular form given in Table 3.

Example 2:
A multi-national company wants to enhance the security of its wireless network. The company plans to buy the next generation firewall for blocking illegal access. For that, the company hire an expert and provide him four next generation firewalls, i.e. U = {u 1 , u 2 , u 3 , u 4 }. The expert have to finalise one of the best next generation firewalls in these four firewalls based on the three given attributes T = { 1 , 2 , 3 }, where 1 represents scan for viruses and malware in allowed collaborative applications, 2 represents identify and control applications sharing the same connection, and 3 represents decrypt outbound SSL. Based on these attributes the expert assigns grades to given four next generation firewalls in the form of stars. A 3-soft set can be obtained from Table 4, where three stars are for 'very good', two stars are for 'good', one star is for 'normal', a hole is for 'poor'.
(Y, (I, T, 4)) Then by Definition 14, the PF5-SS is given as follows: Clearly, the PF5-SS can be described in the tabular form given in Table 5 ( Table 6) Any PFN-SS over U is also called PF(N+1)-SS. E.g. in Example 1, we have PF5-SS but we can also call it PF6-SS. In PF6-SS we suppose that we have 5 grades which are never used in the Example 1.
Here we describe two kinds of containment in PFN-SS, i.e. F-subset and M-subset.

Definition 18:
Example 5: Let PF5-SS over U in tabular form as given in Table 8, which is efficient as we can see that for u 4 ∈ U and 2 ∈ T we have (4, 1.0, 0.0, 0.0) in Table 11.

Applications of the PFN-SS
In this part of the manuscript, we interpreted the algorithm for the real-life problems that are presented in the form of PFN-SS. To show the usefulness of the PFN-SS we present the two examples (selection of corona vaccine and selection of next generation firewall) of DM using PFN-SS.

Algorithm
We describe an algorithm for the selection of the alternative in the environment PFN-SS. We have the following 9 steps of the algorithm. The flow chart of of defined algorithm is given in Figure 2. non-membership pole. 5. Create the score tables for membership pole, neutral membership pole, and nonmembership pole. 6. Add the scores of neutral membership pole, and non-membership pole. 7. Find out the final score by subtracting the sum of the scores of neutral membership pole, and non-membership pole from the score of membership. 8. Observe the greatest score with the highest grades, if it is in k − th row, then we will select option u k , 1 ≤ k ≤ n. 9. The end.

Selection of Corona Vaccine
The world is in the midst of a COVID-19 pandemic. A huge number of people are as yet exposed to coronavirus. It's just the current restrictions that are saving extra people from the deathbed. A vaccine would show our bodies to oppose the infection by preventing us from getting coronavirus or if nothing else making Covid less dangerous. Having a vaccine, alongside better treatments, is the exit strategy. WHO and other medical companies are working on the development of the corona vaccine. Nowadays few medical companies succeed in the development of the corona vaccine. In the following example, we will show that a country can select the best corona vaccine using PFN-SS.

Example 14:
Suppose a country wants to buy a corona vaccine, for this government hires a health expert who will select the best corona vaccine. Let U = {u 1 , u 2 , u 3 , u 4 , u 5 } be a set of five corona vaccines which is under consideration of a health expert to buy and T = { 1 , 2 , 3 } be a set of attributes, where 1 represents price, 2 represents experimental results of vaccines, and 3 represents delivery time. Based on these attributes, the health expert assigns grades to the corona vaccines in the form of stars which is given in Table 1   .   in Section 3. The 5-SS is presented in Table 2 of Section 3 and PF5-SS is presented in Table 3 of Section 3.
The tabular form of the MP is given in Table 25. Next, we create the comparison table for MP which is presented in Table 26. After this the membership score for every corona vaccine with the sum of grades 5 =1 is gotten by subtracting the column sum from the row sum of Table 26. We presented it in Table 27.
Likewise, the tabular form of the nMP is given in Table 28. Next, we create the comparison table for nMP which is presented in Table 29.
After this the neutral membership score for every corona vaccine with the sum of grades 5 =1 is gotten by subtracting the column sum from the row sum of Table 29. We presented it in Table 30.
Similarly, now the tabular form of the NMP is given in Table 31. Next, we create the comparison table for NMP which is presented in Table 32. .     . After this the non-membership score for every corona vaccine with the sum of grades 5 =1 is gotten by subtracting the column sum from the row sum of Table 32. We presented it in Table 33. Finally, we add the scores of neutral membership and non-membership and then subtract it by the score of membership as presented in Table 35. In Table 34, we show the addition of scores of neutral membership and non-membership.
The corona vaccine u 4 has highest grades with the greatest score so the health expert wants to buy the corona vaccine u 4 for his country. He thinks that u 4 is the best corona vaccine among the present 5 corona vaccines which can save more people from the corona.

Selection of a Next Generation Firewall (NGFW)
The firewall refers to a security gateway that protects the computer networks from malicious instructions. There are several kinds of firewalls in the market and every firewall has its benefits and disbenefits. Therefore, how to select a beneficial firewall by seeing its effectiveness and precision is a conventional multi-attribute DM (MADM) problem.
Chances are you have probably heard or read about next generation firewalls, it seems like everyone is talking about what a great innovation it is in the network security market. Traditional firewalls are becoming less capable of adequately protecting enterprise wireless networks with the rapid evolution of applications and threats. Instead of a completely new approach to wireless network security, the traditional firewall has been transformed into next generation firewall to restore application visibility and control and regain the advantage in protecting the enterprise wireless networks. See Figure 3.
In this article, we discussed that a company wants to purchase a next generation firewall whose details are given in Example 2 Section 3. Below in Example 15 we will reconsider Example 2 to get the result.

Example 15:
A multi-national company wants to enhance the security of its wireless network. The company plans to buy the next generation firewall for blocking illegal access. For that, the company hire an expert and provide him four next generation firewalls, i.e. U = {u 1 , u 2 , u 3 , u 4 }. The expert have to finalise one of the best next generation firewalls in these four firewalls based on the three given attributes T = { 1 , 2 , 3 }, where 1 represents scan for viruses and malware in allowed collaborative applications, 2 represents identify and control applications sharing the same connection, and 3 represents decrypt  outbound SSL. Based on these attributes the expert assigns grades to given four next generation firewalls in the form of stars. A 3-soft set can be obtained from Table 36, where three stars are for 'very good', two stars are for 'good', one star is for 'normal', a hole is for 'poor'.
The set Y = {0, 1, 2, 3} can be associated with the stars given in Table 36, where 3 represents  , 2        Clearly, the PF5-SS can be described in the tabular form given in Table 38.
The tabular form of the MP is given in Table 39. Next, we create the comparison table for MP which is presented in Table 40.
After this the membership score for every firewall with the sum of grades 4 =1 is gotten by subtracting the column sum from the row sum of Table 40. We presented it in the Table 38 (Table 41). Likewise, the tabular form of the nMP is given in Table 42. Next, we create the comparison table for nMP which is presented in Table 43.   .  is gotten by subtracting the column sum from the row sum of Table 43. We presented it in Table 44.
Similarly, now the tabular form of the NMP is given in Table 45.
Next we create the comparison table for NMP which is presented in Table 46.
After this the non-membership score for every firewall with the sum of grades   shape of stars which is presented in Table 50. Where Four stars are for 'Excellent', three stars are for 'very good', two stars are for 'good', one star is for 'normal', a hole is for 'poor'.
The set Y = {0, 1, 2, 3, 4} can be associated with the stars given in Table 50, where 4 represents , 3 represents , 2 represents , 1 represents , and 0 represents .  Table 51. An expert link an IFN with each of the grade given by his self. In the result, he will get IF5-SS which is presented in Table 49 (Table 52).
To explain this more we consider the cell in the 1st row and 3rd column, i.e. 3, (0.75, 0.1), where 3 is the grade given by the expert to the candidate u 1 on his working ability, 0.75 be the grade of membership which is given by the expert to the candidate u 1 based on his working ability and 0.1 is the grade of non-membership given to the candidate u 1 by the expert. Now we will use the algorithm defined for IFN-SS by Akram et al. [46] to find which candidate will win the election in the expert's opinion. The tabular form of MP is described in Table 53. We create the comparison table for MP which is presented in Table 54. After this the membership score for every presidential candidate with the sum of grades 3 =1 is gotten by subtracting the column sum from the row sum of Table 54. We presented it in Table 55.
Similarly, now the tabular form of the NMP is given in Table 56. We create the comparison table for NMP which is presented in Table 57. After this the non-membership score for a   is gotten by subtracting the column sum from the row sum of Table 57. We presented it in the Table 58.
Finally, we subtract the score of non-membership by the membership which is described in Table 59.
According to the algorithm defined by Akram et al. [46], the expert found that presidential candidate u 2 has highest grades with the greatest score so he thinks u 2 will the next president of America.     Later the expert realised that in this DM, he forgot to count the grade of neutral membership. Now he wants to add the grade of neutral membership in this DM because he knows that the voters may be categorised into 4 groups, i.e. vote for, abstains, vote against, and refusal of the voting. In the above DM, he has no abstain so now he wants to add this in this DM to get a more accurate and exact result. But he can't add this grade of neutral membership in the environment of IFN-SS. IFN-SS is unable to solve such data. To get the result the expert will use PFN-SS. Reconsider Example 15, The expert will link a PFN to every grade given in Table 51. As a result, he will get PF5-SS which is presented in Table 60.
To explain this more we consider the cell in the 1st row and 3rd column, i.e. 3, (0.75, 0.01, 0.1), where 3 is the grade given by the expert to the candidate u 1 on his working ability, 0.75 be the grade of membership which is given by the expert to the candidate u 1 based on his working ability, 0.01 is the grade of neutral membership, and 0.1 is the grade of non-membership given to the candidate u 1 by the expert. Now we will use the algorithm defined in Section 4 of this article to find that which candidate will win the election in the expert's opinion. The tabular form of MP is described in Table 61. We create the comparison table for MP which is presented in Table 62. After this the membership score for every presidential candidate with the sum of grades  Likewise, the tabular form of the nMP is given in Table 64.We create the comparison table for nMP which is presented in Table 65. After this the neutral membership score for every presidential candidate with the sum of grades 3 =1 is gotten by subtracting the column sum from the row sum of Table 65. We presented it in Table 66.
Similarly, now the tabular form of the NMP is given in Table 67. We create the comparison table for NMP which is presented in Table 68. After this the non-membership score for a presidential candidate with the sum of grades 3 =1 is gotten by subtracting the column sum from the row sum of Table 68. We presented it in Table 69.
Finally, we add the scores of neutral membership and non-membership and then subtract it by the score of membership as presented in Table 63. In Table 70, we show the addition of scores of neutral membership and non-membership.

Conclusion and Future Study
It is observed that one of the essential concept of neutral grade is missing in the IFN-SS theory. Concept of neutrality grade can be observed in the situation when we encounter human views including more answers of type: yes, abstain, no, refusal. To overcome this issue, we interpreted the idea of PFN-SS by merging PFSs and N-SSs in this manuscript. We defined some relevant properties of PFN-SS such as a compliment, restricted intersection and union, extended intersection, and union, M-subset, and F-subset of PFN-SS. We also linked PFSSs with PFN-SS by using a threshold. Further, we gave numerical examples to explain these notions. We described an algorithm to deal with PFN-SS data. To show the advantage and usefulness of the defined technique, we used two examples which are 1. Selection of corona vaccine 2. Selection of next generation firewall from real life by utilising PFN-SS data. In the selection of the corona vaccine we have seen that the corona vaccine u 4 was the best vaccine to purchase according to the health expert. Similarly in the selection of a company for a job, we have seen that the company u 1 was the best company for a job according to the expert. After this, we did a comparison of our work with IFN-SS by using Example 16 given in Section 5 in which we have seen that how PFN-SS change the result which an expert got by using IFN-SS data. From that example, we got the results that our defined algorithm in Section 4 is more general than the algorithm defined by Akram et al. [46]. One can solve IFN-SS data by our proposed algorithm by taking the neutral membership zero. By our proposed algorithm one can also solve the FN-SS by taking neutral membership and non-membership zero. This showed that our initiated method is more general and suitable than the intuitionistic fuzzy N-SS (IFN-SS), fuzzy N-SS (FN-SS), and N-SS.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Data Statement
The data used in this manuscript is artificial and anyone can use it without prior permission by just citing this article. He has more than 180 international publications to his credit and he has also produced more than 46 MS students and 6 Ph.D. students.