Approach to developing expert system

Approach to developing expert system

At the market today, we can find many tools for creating expert systems. These systems can be developed in a programmable environment through tools of type C + +, Visual Basic or some other programs which are related to development of expert systems. However, today are developed specialized tools for creating expert systems which allow a high degree of automation in process of developing expert systems. There are called expert system shells. From the standpoint of this research it was carried out choice of expert system shell from the aspect of next four criteria [43-45, Personal communication with group for consulting from London South Bank University, Business, Computing & Information Management, 2011):

- programmability,

- comprehensiveness,

- universality,

- price.

During the election, it was analyzed 58 shells. All information about shells are available on the Internet [2], and classified in a group of commercial shell. Detailed analyses were conducted separately for each tool through analyzing belonging site. For evaluation on the basis of the criteria it was adopted the scale of 1 to 5 where 1 is unsatisfactory grade. According to defined criteria as a most distinguished tool for the needs of the research was adopted ACQUIRE shell. That tool is non programming oriented and it has affordable price. This is a tool that supports the work of the Windows operational environment. It has possibility to develop all elements of expert system and supports forward, backward and combined chaining. For the presentation of knowledge it can be used production rules, the action table, or combined techniques. During a process of developing expert system, the role of engineer for knowledge took up first author, and the role of one expert took up second author. Also, as sources of knowledge were used following:

- experience from eleven prestigious organizations in the world of field of quality management systems, business excellence and organizational performance [46],

- guidelines from standards for improving organizational performance [47 ],

- best practices from auditing of ISO 9001 oriented system [48],

- experience and practice of organizations that participated in the competition for the Oscar of quality award [49],

- theory and principles of TQM [50],

- experiences that are listed in [51] and indicate the path to business excellence.

The expert systems are included and knowledge gained through many concrete practical projects of quality management systems implementation, and many training on that topic. That knowledge is next:

- knowledge that are specific to certain companies,

- knowledge derived from specific experiences and on specific way of solving problem,

- knowledge of those that are best for certain jobs and are passed special training,

- knowledge of those that is proven in practice for the specific job and similar.

For the purposes of this research, expert system was develop for modules 5 (management responsibility) and module 8 (measurement, analyses and improvement) of ISO 9001 standard. The reason for that is that these areas have the greatest importance in achieving business excellence [1] and therefore they should be considerate from the standpoint of improvement. Also, another reason is that module 8 has requirements that are oriented to the improvement and that is essence and priority.

The idea of this research is to make the integration of decision support systems (DSS) which is operate on first level of experimental data, and expert system. That is modern approach of integration a number of tools with the aim of acquiring a larger volume of better knowledge [52] and make system with higher level of intelligence. Today trends are integration expert systems and traditional decision support systems which as output give data and information [53].

Integration of expert systems and decision support system can be achieved in two ways [54]. The purpose of this research is to use model which is present on figure 5. based on the collection and analysis of data obtained at the output of the decision support system and it provide important information like one of inputs for expert system and its knowledge base. This is the model which is completely compatible with previous remarked analogy with human body. This two approach stay in base of this analogy integrative model for improvement business process performance.

For the purposes of this research, we developed a decision support system in the MS Access, Select Query Language and Visual Basic environment.

Applying Pareto method and rules of 70/30 it can be identified area which is crucial from the standpoint of improvement. Also, this system like support for making decision provides written presentation of nonconformities which can be shown as experience of other companies. That could be use like important data for the definition of knowledge in expert system. In addition, this system provides, and comparative analysis with the period of the four years before, which also has significance for the definition of knowledge in the expert system.

Connection between data from the first and data from the second level was achieved through the introduction of the concept of "Degree of readiness (Si)" in achieving business excellence, in accordance with the following expression:

Si = Nz [%] * Kz , i=1,2,...,26

where:

Si        Degree of readiness for all type of organizations for all requests of ISO 9001

Nz      Power of a standard clause in terms of percentage. Nz = ƒ (number of nonconformities from experimental database)

Kz      Coefficient of significance for achieving business excellence

That degree is applies to every single request of ISO 9001 and showing the willingness or the ability of organizations (both manufacturing and service sector) to attain business excellence in some areas. To find this degree, we are using method Analytic Hierarchy Process (AHP) and corresponding software Expert Choice. 

It is important to emphasize this because it was used and it is very important during definition of preventive measures in terms of defining their priorities and "power". Also, “power” of prevention was related with number of nonconformities in particular area. That means, larger number of nonconformities, or larger number of experience, make possibilities for defining more effective and efficient preventive action like output of expert system.

Through application of Pareto method, based on coefficient of significance following requests were identified as the most significant for achieving business excellence:

requests - 821, 823, 85, 84, 54, 824, 56, 53, 71, 41, 51, 72, 55

At the same time, this is important areas, and have high level of priority for improvement from the standpoint of achieving business excellence and it is very important for defining preventive action of expert system and intensity of that action. If we take a look at the list of "Coefficients of significance" for business excellence achieving, especially the most important ones and perform comparison with the list of variables and their significance in terms of: Business Process Reengineering (BPR), manufacturing strategy, benchmarking and performance measurement, being the result of the appreciated research [55] and [56] it may be found significant intercompatibility.

The concerned compatibility is especially reflected in the following variables, evaluated in the relative research as highly significant for the following four projects, i.e.: customer satisfaction, quality, employee satisfaction and personal growth, customer adaptability, identification of top managers with BPR goals, strong process orientation, results orientation, direct customer cooperation. On the other hand, the above mentioned four areas are considered as highly important for any market-oriented organization, thence it can be concluded that organizations by strengthening their capacities in areas of presented "Coefficients of significance" (especially the most important ones), are not only strengthened in terms of the business excellence achieving as per European Award model, but also in the stated four areas.

But some of these areas are much more important then other. Because that, the research was further elaborated in order to indicate most important area for improvement and area where should be focus attention and where should be provide very intensive action in order to achieve best organizational condition and results. This research was conduct from the standpoint of occurrence of nonconformities in all type of organisation regardless of they size or type (both for manufacturing and service organisation). Parallel the Pareto method (70/30) was carried out in that direction and based on that, it was identified next areas:

requests - 56, 75, 62, 822, 74, 76, 54, 72, 85, 821, 55, 63

Now we are search for common requests (area) that are most important and where should be oriented focus and where should be provide extensively action in terms of achieving business excellence regardless of type or size of organization. And they are:

1. 821 – customer satisfaction,

2. 72 - customer related processes,

3. 54 – planning,

4. 85 – continual improvement,

5. 56 – management review and

6. 55 – responsibility, authority and communication.

This area is most important for defining output of expert system and for defining intensity of action for improvement.

Objects were defined during the process of expert system developing. That were depend of problem which should be solved, base on ISO 9001 oriented check list an based on experience which can be find on DSS output. Base on results of DSS system, it is defined value of the object and relation between them.

At the end, after starting the program, in a short time, system introduce user in a set of dialog boxes.

Data obtained from this report, user can use and implement knowledge that an expert system produces. However, users can improve performance of an organization in the field where such as performance are on lower level. Also, it can be improvement performances of other, non critical, area and can be reach level of business excellence.

This expert system was developed in three iterative steps. Each of them resulted of the improvement, for example improvement of the definition of objects, set the input data, the relation between objects depending on the priorities of execution and more.

The expert system was implemented and tested in practical, real conditions in the organization that has a clear commitment to participate in the competition for the European Award for Business Excellence, also providing important measures in that direction. Evaluation was done on the basis of technical and ergonomic characteristics based on guidelines in standards ISO/IEC 9126/1:2001 for evaluation quality of software.

Software was evaluated positive in terms of technical characteristics and in terms of ergonomic. In

this sense, product has small time of response, it is compatible with most used operating system, it has an excellent user’s oriented interface, and it has easy data entry and a good view of the output, installation is simple and the software is very competitive. Also, in this sense, within the organization, it was carried out the reorganization of the priority areas from the viewpoint of improvement, implemented preventive measures for the potentially unstable areas and also applied the measures for the improvement (offered by this system) leading to business excellence achieving.