Ўзбекистон республикаси ахборот технологиялари ва коммуникацияларини ривожлантириш вазирлиги муҳаммад ал-хоразмий номидаги

 
Адабиётлар: 
1. И.А.Каримов маърузаси. // “Халқ сўзи”. 2014 йил 18 январь. 
2. Қ.Х.Абдураҳмонов ва бошқ. “Персонални бошқариш”. Ўқув қўлланма. Т.: “Меҳнат” 
2004 й. 
3. Развитие организации и HR-менеджмент: Учеб. пособ. М.: “Независимая фирма 
класс”, 2004 г. 
 
 
 
 
 
 
 
 

---

97 
SOME APPROACHES CONSTRUCTION PRODUCTION SYSTEMS INFERENCE 
FOR ELECTRONIC INFORMATION RESOURCES 
N.O.Rahimov (TUIT) 
U.Murtazayeva (Samarkand branch of TUIT, assistant teacher) 
 
The paper deals with some approaches of building the productive inference systems for 
electronic information resources. The analysis of the construction of acquiring knowledge and 
approach of building production systems for electronic information resources.
The rapid growth of data volumes requires innovative approaches to address them collecting 
and reporting tasks in the system storage of electronic information resources (EIR), as well as 
their subsequent processing and analysis by means of information-computing systems (ICS). 
Processing involves the solution of related search tasks, pre-processing, structuring and analysis 
of data, as well as their categorization (classification) in the information resources (IR). 
For data analysis, a number of approaches EIR proposed in which procedures are 
implemented at different processing stages of the analysis. The quality of the data analysis and 
subsequent decision tasks (evaluation of current and projected conditions, decision support, 
management, and others.) Depends on the quality of the contents of both data and knowledge 
base, i.e. by EIR in general [1]. 
If we do a comparative analysis of knowledge representation, then it can be seen that 
production models have the following advantages over models of semantic networks and frame 
model [2]: 
- A significant portion of human knowledge can be represented in the form of products; 
- Simplicity of construction and use; 
- High interpretability; 
- The availability of advanced inference engine. 
Thus, the production model of knowledge representation is the most effective and popular 
for use with other existing models of knowledge representation. 
In general, a production model can be represented as follows [3]: 
N = < A, U, C, I, R> 
N - the name of products; A - the scope of the product; U - the condition of the applicability 
of the product; C - core products; I - post conditions products, actualizing for positive product 
sales; R - comment, informal explanation (justification) products, the introduction to the 
knowledge base. 
Knowledge processing systems that use a production model are called “production system”, 
based on the rules of inference Post, presented in the form of schemes 
 t
1
,t
2
,…,t
n 
 
t 
where t
1
, t
2
, ..., t
n
- this parcel, and t -conclusion. 
This system is the basis for the conclusion of direct rules (modus ponens - m.p.) form 
(scheme) 
А, А → В 
B 
those. if A and A → B - are true, then the two formulas A and A → B can be obtained in a 
new formula, where A, B - value respectively assumptions and conclusions).(In production 
systems are also used and generally reverse conclusion (modus tolens - m.t. 
B, А→В
A
These rules are the foundation of the propositional calculus and predicate, in which the 
following inference rules are used: substitution and conclusions. substitution rules provide for a 
variable wherever it occurs in the formula, the substitution of one and the same formula.The 

---

98 
structure of production system includes a rules database (Product), global database and 
management system.When using such models in systems based on knowledge, it is possible to: 
• Use simple and precise knowledge of the use of the mechanism; 
• knowledge representation with high uniformity, described by a single syntax. 
This form of knowledge representation has the following advantages: Natural man expert in 
many cases expresses his knowledge is in the form of regulations; Modularity: each rule is a 
relatively independent fragments of knowledge that makes it easier to debug and modify the 
knowledge base; Transparency: the convenience of the explanation of the withdrawal process 
solutions. 
However, the major drawbacks allocate 3 production systems: 
1. The difficulty of drawing up a production rule corresponding element of knowledge. We 
need to treat the area has already been sufficiently studied and established good primitives and to 
the level of detail was not too detailed, or would be required to have one rule for every situation. 
2. The difficulty of recording rules. A single IF-THEN format record results in cumbersome 
expressions on the left side and repeat the same parcels in similar situations; with its help it is 
difficult to express complex rules. 
3. Increase in the number of rules in the knowledge base. The sharp slowdown of inference 
with an increase in the number of rules in the knowledge base, which is unacceptable for systems 
operating in real time. Also, the accumulation of a sufficiently large number (several hundreds) 
of productions they begin to contradict each other. 
Let us turn to the consideration of approaches to address the shortcomings of production 
systems. In addition to the base size of the rules it is important to the structure of production 
rules themselves. For the construction of the rules must be compatible: 
- Use the minimum sufficient set of conditions when determining the production rule; 
- Avoid contradicting production rules; 
- Design rules, based on the structure inherent in the subject area.
Systems based on rules that allow you to combine a group of related pieces of knowledge. 
Each production rule can be used independently of the other. This independence makes the base 
of production rules semantically modular, i.e. group information does not affect each other. This 
allows us to develop a knowledge base. 
Construction of production base of knowledge and grouping rules in the group is performed 
manually by experts, and the structuring of the rules in the knowledge base - automatically. 
In [3] developed a mathematical model, which is a tabular model that integrate all the 
entities and dependencies, represented in the knowledge base. Another problem - the dimension 
of the rule base. Since the dimension of a system of rules determined by the number of object 
parameters in a need of controlling parameters. In [5] reduction rule base dimension offers to 
carry out the following two ways. Firstly, to solve many practical problems diagnosing the state 
of objects can be enough to use the information only on the presence or absence of certain signs, 
which are attributes of the state. Second, to solve the problem, in many cases, acceptable results 
can be obtained by using an aggregate for some information about the rules of the object 
parameters. 
The growth of contradictions production model may be limited by introducing arrangements 
exceptions and returns. 
To resolve the contradictions in the knowledge base developed a series of mathematical 
algorithms, such as automatic theorem proving, busting mechanism to return and others [4]. 
Resolving conflicts at the stage in which selects and activates one of the acceptable 
products, the following strategies apply: 
- Refraction to prevent loops: after activating the rules, it cannot be used again until you 
change the contents of working memory. 
- The novelty of search allows you to focus on one line of reasoning: preference for rules, 
provided that there are facts that are added to working memory last. 
- Specificity prefers more specific rules before more general: one rule is more specific 

---

99 
(concrete) than the other, if it contains more facts in the conditional part. 
Currently, there is no uniform method of building knowledge bases, which causes some 
difficulties in modeling systems. A combination of approaches considered above allows you to 
compensate for the shortcomings production systems and highly specialized to solve the 
problem, but in the development of a universal method of creating a knowledge base of varying 
complexity and applicability remains open.