Sergeeva Elena

Abstract

Theme:"Computer subsystem of the calendar planning in a build enterprise”

Russian Version

One of the major tasks of planning on a build enterprise is development of the planned schedule building works on the set technology. A good planned schedule lets to use possibilities of productions of enterprise in the best way. It sets the concrete terms of works.

It is very important, that a calendar plan to meet a production environment and also all external and internal limitations. Quality of calendar plan is determined precision and timeliness of input data information which it leans against. Efficiency and correctness of plans that are used for control of processes of productions directly affects industrial enterprise activity.

Therefore a problem of development of a computer system of a planned scheduling in a build enterprise is actual. This system will allow quick and effective making of a calendar plan for an enterprise that will increase its productivity and building quality.

The purpose of my master's degree work is development of a computer system of a planned schedule in a build enterprise. This system will be built on the base of modern methods and techniques.

I will endeavor to avoid in my master’s work all those failings that are in the existent systems.

There are a great number of methods for planned schedule making:

1. Method of linear plan-graph of the Gant

2. Model of calendar plan as cyclorama

3. Network planning

4. Planned scheduling on a base of artificial intelligence (expert system)

Each of these methods has both the dignities and failings.

The linear graphs (graph of the Gant) in the form of ribbon graph are most easy widely applied in building.

However, the linear graph for building large and difficult objects cannot achieve title goal, as it does not allow planning numerous intercommunications of elements of production, to choose the optimum variant of execution of all program to use reserves and to correct operatively the graph during building.

Model of calendar plan as cyclorama also got wide distribution in practice of building.

Cyclorama is especially comfortable at erection of the same types buildings, because enables evidently to reflect development of stream in time and in space.

Shortcomings of the linear calendar plans are removed at the use network models. It allows not only to expect the basic parameters of the graph, determining execution of the program time on the whole and began terms completions of associate works in composition a complex but also analyze the graph, expose backlogs and use them for his improvement (optimizations).

A network model of planned schedule is most successful for the reflection of probabilistic productions process. It allows to represent complicate object building in more evident form.

Approach of the expert systems is applicable to a planned scheduling, because it assists coordination of universal knowledge with the special information, provided by guidance of building, information of a build control and any other information, passing in the structure of the system.

An intellectual expert system (based on knowledge) is able to decide problems using knowledge, coded and got from experts.

In the master's degree work I want to build an expert system of the planning schedule for a build production. This system ware taken into account:

1. The alternative arbitrary complication network model is formulated as a linear rule sequence that makes the base of knowledges of subject domain. It can be filled up or modified in the process of researches and on the basis of the real systems exploitation.

2. The rules, originally produced at the expert system creation may be imperfect. Therefore, the results should be analyzed carefully to specify and sift form inadequate rules, i.e. to reflect actual knowledge of expert.

3. Only a little number of rules representing the state and specific of the concrete object adds usually to the general base of knowledge. It differs substantially the examined approach from the graph methods, there a change or addition of model with «cross» connections requires principle and rather significant processing.

4. Because of nondeterministic nature of the planned problem it is necessary to capture data in the most favorable time, to guarantee the reception of the grounded, realized decisions at the construction of time-tables.

5. It there are not the rules for the concrete situation in the base of knowleges the system must be able to make a decision on the basis of the present information. The system must be able not only to correct a situation in time but also to look over the base of knowledges as necessary by additing, deleting or correcting of maintenance.

6. Some professional language should used for a formalization of descriptions It is the usual method of technology definition for user.

7. The use of the developed logical methods systems allows to check up models admission in the automated mode effectively.

8. Inability to forecast the future and necessity to make decisions after every event may reduce efficiency of the system; therefore it must be able to prognoses.

Beginning