Every enterprise competently requires in use of resources they have in the modern conditions of the economic crisis. With a purpose of a rational distribution of these resources, it´s necessary to detail the production activity.
The enterprise competitiveness and its adaptation to an unfavorable crisis´s influence depends from a prognosis´s quality.
One of the most important task for Opened Joint-Stock Company «Samarskiy fatindustrial complex» is a forecasting of sales volume and to form plans of the margarine production.
The researches presented in this work were conducted by an order of an enterprise within the limits of making production political strategy.
The researches conducted beforehand exposed that the consumption of margarine has expressed seasonal character.
It leads to that, the sales volume forecasting task can be solved by means of a full model (if not to count a seasonal factor), so as a partly seasonal.
There is a lack in the first method in which is impossible to consider the production´s features, which connected, for example, with a seasonal scheduled repair of equipments and its setting.
This model has fictitious connections between the seasons which are practically unimportant. However, an advantage of this method has a possibility of using a big statistical material in a short time.
Considering all these features, the second method allows to build a general model seasonally more realistically.
However, to get a representative sample it is necessary to use statistical material, beginning from 2003 year.
It is a certain lack, as the situation in Samarskiy region was changing for so long period.
Here are following results:
1. The full model of sales volume of margarine is built disregarding a seasonal factor.
2. It is built the partly seasonal model.
3. It is estimated a model quality by means of a determination factor and a consent criterion of hi-squire of Pirson and it is lead comparative analysis of its correctness.
4. It is concluded, that a partly seasonal model corresponds actual sales volumes.
All researches in this work are made by the author personally, and the results used by the enterprise in its activity.
The work was submitted to electronic scientific conference «Mathematical modeling», came to the editorial office on 27.09.2010.