Samoylov A., Semenenko N.

Southern Federal University, Russia

Application of decision support system based on indistinct logic for choice mobile phones

The market of mobile devices is actively developing now. Updated mobile phones are increasingly available. And also advanced companies realize a sharp increase of mobile device’s functionality. Thus, phone selection is difficult for the average user, because the phone must meet the basic requirements of user. There is great number of proposals in the market of mobile phones. This fact makes complexity of finding a suitable phone. And sometimes the user can not explore all proposals physically. Because of this, the user must buy a bad mobile phone, which may not suit to him personally. The problem of development a specialized information system for choice a mobile phone that meets the requirements of a particular user is very important. We offer to use the method for decision this problem. It is based on the decision support system.

There are three classes of decision support systems. Kind of class depends on the complexity of the task and the field of application.

The systems of the first class have the highest functionality. They are designed for use in the public administration of the highest level (the presidential administration, ministries) and control department of large companies (board of directors of a corporation) in the planning of large complex target programs to explain decisions of inclusion in the various political, social or economic activities and the distribution of resources between based on an assessment of their influence on the achievement of the main objectives of the program.

The decision support systems of the second class are systems for individual use, their database is formed directly by the user. They are intended for use by government officials of medium rank, and managers of small and medium-sized firms for operational management tasks.

The decision support systems of the third-class are systems for personal use, which are adapted to the experience of the user. They are designed to address common application tasks of systems analysis and management (for example, the choice of the subject of credit, appointment to a post etc.). These systems provide a solution to the present problem on the basis of the results of the practical use of solutions of the same problem, taken in the past. In addition, this class of systems is intended for use in commercial establishments that sell expensive durable goods. [1].

Analysis of all existing classes of the decision support systems has shown that the best way for choice of mobile phone is using decision support system of the third class. It can provide the best combination of price and quality of decisions. And it will be a system for individual using for everybody. The introduction of indistinct logic in the decision support system allows the use of natural language. Based on the previous experience of studying decision support systems and software, which realize the possibility of their using, we concluded that it is appropriate to use a software tool Conclution, which will create an information system based on indistinct logic to select the desired mobile phone. Let's look at the process of creating this system. The first step is to determine the name of linguistic variables and their number, based on the subject field. For our case of buying a mobile phone such variables can be: price, company, the manufacturer and availability Wi-Fi.

Fig. 1. The introduction of linguistic variables

Next, you must determine the terms for every given linguistic variable. It should be noted that the meanings ​​of the variables is words of natural language and called terms in the indistinct logic. Therefore, we can choose the terms for each linguistic variable. There are «low», «medium» and «high» in a variable «Price». There are «Nokia», «Samsung» and «HTC» in the variable «Firm».  There are «Finland», «South Korea», and «Taiwan» in the manufacturer. There are «yes» and «no» in the availability Wi-Fi. The exact meanings ​​of the input variables are converted to the meanings of linguistic variables through the application of certain provisions of the theory of indistinct sets, namely with certain membership functions. Next, we find the numerical meaning or range of meanings ​​for each term of linguistic variable. Now it is necessary to consider the indistinct rules that can describe the control strategy for choice of a mobile phone.

Fig. 2. The construction of the membership function

 

We need to determine the grade of belonging of an antecedent of rule in the logical inference. There are MIN (…) and MAX (…) in the indistinct logic. The MIN (...) calculates the minimum meaning of the degree of belonging. And the MAX (…) calculates the maximum meaning of the degree of belonging. The application of the particular operator depends how the parcels in the rule are connected. We have to apply the operator MIN (...), if we use "AND".  We have to apply the operator MAX (...), if we use "OR".

Work out the above procedure for each rule from the base of indistinct rules:

Ïk,r = MIN ËÏi (ËÏi is considered in the rule and ËÏi is Òi,l, where l is 1..mi),

Ïk,r is the rule,

k = 1..p, p is number of rules, r  is number of solutions, that corresponds to this rule;

ËÏi  is the linguistic variable, i = 1..n, n is number of  the linguistic variables.

Òi,j – j- term of  i - the linguistic variable, j = 1..mi, mi – number of  i - the linguistic variable.

Fig. 3. Creating a database of indistinct rules

 

Let’s write the results in a software environment Conclusion. Thus, we have developed a system by which the user can enter the data and functionality of the phone. The system prompts to client to buy the best option.

We tested the system during the month. We concluded that the system copes with the task of selecting a mobile phone perfectly thanks to the knowledge of the functional characteristics of the phone. However, resources are needed for the entry of all linguistic variables related to mobile phones, their terms and creating a database of the indistinct rules for decision. Therefore, we decided to upgrade the system, we used the web – Technology. It will allow users to work with the Internet. Every internet users will be able to visit the web - site and add the linguistic variables, their terms and create an indistinct rule specifically for themselves. It will extend the system and increase the quality of decision.

 

References:

1. Totsenko B. Decision support systems - your tool for the right choice - the magazine "Computerra" ¹ 34 dated September 1, 1998

2. Grinyaev S. Indistinct logic in control systems - the magazine "Computerra" ¹ 38 dated October 8, 2001

3. Gajzler M., K. Zienkiewicz The use of indistinct logic to decide in expert systems - Publishing of Poznan University of Technology: Poznan 2004,

ISSN 1642-9303 No5-2004