Українська   Русский
DonNTU   Masters' portal

Summary of the topic of graduation work

Содержание

Introduction

The main goal of any store is to meet the needs of consumers as much as possible, as well as to increase the number of sales, this in turn leads to maximizing the store's profit. To achieve this goal, first of all, you need to study preferences, customer needs, and the main factors that influence price formation. Many online stores today face problems attracting buyers, this may be due to the wrong set price for the product.

Price formation is a complex multi-faceted process. The information required to set the price should be to learn to perform. The lack of information, as well as its excess, makes it difficult to solve the problem.

This paper analyzes the methods of forecasting demand for the formation of prices for goods. Factors are investigated, influencing the prices of goods, the criterion of optimality of the price of goods is formulated, the model of the intellectual system is proposed calculating the optimal price for products based on the sales history and prices of competitors.

1. Actuality of theme

At the moment, in order to set the optimal price for a product, it is necessary to forecast the demand for goods and services. By means forecasting can minimize risks, costs, and build an accurate plan for setting the price for the product that will be relevant among consumers, this will lead to profit, as well as strengthen the organization's position in the market of goods and services.

Based on the analysis and identified shortcomings of existing developments in this work, a study is conducted aimed at solving the problem of forecasting demand and setting the price of a product using forecasting methods.

Price is an economic category that means the amount of money that the seller wants to sell for and the buyer is willing to pay purchased goods. It contains almost all economic relations in society [3].

All firms and stores are faced with the task of setting prices for their products or services. In order to sell your product or services on the market, you must set prices for them that would be acceptable to buyers, otherwise they can not be successfully sold on market. Therefore, the store must choose the correct pricing policy.

Price formation is a complex multi faceted process [7]. The pricing process is shown in figure 1.

Figure 1 – The pricing process

Many online stores today face problems attracting customers, this may be due to the wrong set price for the product.In today's time there are so many online stores (of konkurentov). When forming prices, it is necessary to take into account the prices of competitors.

At the moment, the most advanced technology for obtaining statistical data of generated prices in an online store is based on demand consumer's. That is, using the collected information, you can analyze the behavior of customers, namely, to determine which product is the most popular or Vice versa is not in demand, to understand whether the buyer is satisfied with the price of the product, and in accordance with the received data to change the prices of goods, which in turn the queue will allow you to change the buyer's demand and then analyze again how the buyer's behavior has changed.

2. The purpose and objectives of the study

The main goal of any store is to get the maximum profit, and this can be achieved by increasing the number of sales. For achievement to achieve this goal, first of all, it is necessary to study the preferences, needs of buyers and the main factors that affect the formation of prices, based on how to develop a model of an information system that generates optimal prices for online store products.

In order to develop such a system, it is necessary to formulate the criteria by which the price of a product is formed, and to investigate the methods that they can be used in such a system and develop a model for price formation.

Let's consider an online store that presents a certain product range T1...Tn. Each product must determine the price of Ц1 ... Цn., to maximize the profit from sales and sell the maximum possible number of products

where

n – number of product types,

m – number of products sold,

Пi – sales of the i-th type of product,

Рi – purchase price of the i-th type of product.

Sales can be calculated using the formula:

где

Кi – sales of the i-th type of product,

Цi – purchase price of the i-th type of product.

3.Alternative methods for solving the problem

In order to set the optimal price for goods, you need to predict demand..

There are such forecasting methods:

  1. simple average Method;
  2. moving average Method;
  3. weighted average Method;
  4. exponential smoothing Method;
  5. Holt-winters Method;
  6. Autoregression Method;
  7. Neural networks, genetic algorithms.

Next, let's look at each forecasting method separately.

Simple average method

It is the simplest of these methods, using calculations based on the formula simple average. The forecast of the product price for the next period is calculated using this method as the arithmetic average of price indicators for all previous periods.

The disadvantage of this method is its high conservativeness – outdated information previous sales will prevent the latest demand trends from showing up at this price.

Moving average method

The moving average method responds more quickly to price changes. The calculation is made in this case not based on data for the entire period of observation, but for several recent periods.

An interesting variation of the method is the calculation of the moving average for certain seasons. Such method may be suitable for products that have a pronounced seasonality[4].

The method of exponential smoothing

Unfortunately, the above methods of calculation on average allow you to get only very approximate Results. For a more accurate forecast, you can achieve exponential smoothing using models and exponential smoothing with the trend.

In the first method, the last sales forecast is adjusted based on the forecast error made in the last one period. The second method of calculation (also called the double exponential smoothing method) takes into account data with trends & ndash; due to this, this method can be used even for medium-term forecasting.

Holt-winters method

Many products tend to increase or decrease sales, especially when they are first produced or when they appear competing goods. For some products, seasonal changes in the level of sales are significant, so for the forecast of product prices it is advisable to take into account the specific nature of the trend and seasonal fluctuations. Based on the Holt winters model) I created my own predictive model that takes into account the exponential trend and additive seasonality.

To get a demand forecast in this method, you need to select three parameters correctly. To do this, you can use as special algorithms, and be limited to a simple search.

The Method Of Autoregression

If you want to get even better forecasts, you can use the autoregression models. This technique allows you to conduct a very detailed analysis of the available data, identify any trends and weed out random influences. However, in contrast to using the previous methods, selecting a set of parameters will require a lot of effort and time from the user.

Forecasting using the autoregression model is based on previous sales values. The word autoregression means dependence of the subsequent sales value on previous sales. The dependence in the case of autoregression is assumed to be linear, then there is a forecast that represents the amount of sales for the previous days with some coefficients that are constant and define parameters of the autoregression model. How many days (periods in General) of such sales from the past will we take to trying to predict future sales at a set price is called the order of the autoregression model [5].

Neural networks, genetic algorithms

It should be noted that the more complex forecasting methods are used, the more difficult their practical application is and the higher probability of errors. Analysis of huge amounts of information, selection of optimal parameters, operational accounting of market fluctuations – all this is sometimes at the limit of human capabilities. The most promising solution to this problem is using neural network algorithms.

In this method, a special program after preliminary training is able to independently find the best solution – when in this case, the user does not need to delve into all the wisdom of the theories used. In addition, neural networks are able to take into account hidden trends and create a reliable forecast in such an unstable situation, where previously forecasting was considered impossible at all [2].

4. Description of the task of improving image quality

In the process of studying existing methods, it was decided to use the neural network method and genetic algorithms.

Dynamic pricing can be organized using artificial neural networks. The training data for the level of demand, depending on the day of the week and the time of day, they are taken for the previous period. The main advantage of a neural network is the ability to learn and get data yourself. Unlike traditional demand models, models built on neural networks do not any preliminary assumptions about the relationship between different factors. Most likely, they will learn these relationships from the data itself. They can derive a value from complex or inaccurate data and can be used to model relationships that are too complex, to be noticed by people or computer equipment. This ability of neural networks makes them a good candidate for modeling demand in dynamic pricing.

Neural networks allow you to solve problems that traditional methods can not cope with, they are able to successfully solve problems based on incomplete, noisy, distorted information [1]..

As a method for optimizing the dynamic pricing problem based on an artificial neural network, it is proposed to use evolutionary algorithm. They use the concept of natural selection and random changes in evolution that allows you to find the best solution to the problem.

Let's consider the scheme of dynamic pricing using a neural network (Fig. 2).

Figure 2-Diagram of dynamic pricing using a neural network

In the course of its activity, the company purchases goods from suppliers at wholesale prices and sells them to the public at retail prices. At the same time the gross income of the enterprise is formed, which is determined by the revenue from the sale of goods and services, minus the cost of paying the cost received from suppliers goods'. The company strives to maximize its net profit, which, under other fixed conditions, including tax rates, depends on values of trade margins on goods. There are other factors, such as customer demand, product competitiveness, and turnover rates, which also affect the amount of profit.

The neural network receives data from various sources (prices from competitors, the level of demand, the price of wholesale purchase of goods) and analyzes them. the optimal price for the product.

5. Factors that affect product prices

In a market economy, pricing in foreign trade, as well as in the domestic market, is influenced by a specific market situations. By their nature, level, and scope, they can be divided into the five groups listed below [6].

  1. General economic rules that apply regardless of the type of product and the specific conditions for its production and sale. These include:
    1. economic cycle;
    2. state of aggregate supply and demand;
    3. inflation.
  2. Specifically-economic, determined by the characteristics of this product, the conditions of its production and sale. These include:
    1. costs;
    2. profit;
    3. taxes and fees;
    4. supply and demand for this product or service, taking into account interchangeability.
  3. Consumer properties:
    1. quality;
    2. reliability;
    3. appearance;
    4. prestige.
  4. Specific, valid only for certain types of goods and services:
    1. seasonality;
    2. operating costs;
    3. completeness;
    4. guarantees and terms of service.
  5. Special, related to the operation of special mechanisms and economic instruments:
    1. state regulation;
    2. exchange rate.

Conclusion

In the course of the work, various methods of forming and forecasting prices for goods were studied. It was decided to create a system using the Costs plus method and the forecasting Neural networks. The simple average method will be used to predict the demand for a product. With this system, the optimal price for products will be formed, taking into account the prices of competitors, preferences and needs buyers, as well as taking into account the season, the life cycle of the product.

While writing this abstract, the master 's work has not yet been completed. Approximate completion date: June 2020. The full text of the work and materials on the topic can be obtained from the author or his manager after the specified date.

References

  1. Использование нейронных сетей для динамического ценообразования. [Электронный ресурс] – Режим доступа: URL: http://edrj.ru/article/19-04-16
  2. Нейронные сети: на пороге будущего. [Электронный ресурс] – Режим доступа: URL: http://www.neuroproject.ru/articles_dak_nn.php
  3. Ценообразование. [Электронный ресурс] – Режим доступа: URL: https://ru.wikipedia.org/wiki/ Ценообразование
  4. Методы прогнозирования цен. [Электронный ресурс] – Режим доступа: URL: https://works.doklad.ru/view/LqjyMIaIGtA/all.html
  5. Прогнозирование. [Электронный ресурс] – Режим доступа: URL: https://fnow.ru/articles/prognozirovanie-sprosa
  6. Факторы, влияющие на формирование цен на мировом рынке. [Электронный ресурс] – Режим доступа: URL: https://www.inventech.ru/lib/cost/cost-0054/
  7. Методы формирования цены: Процесс ценообразования. [Электронный ресурс] – Режим доступа: URL: http://ww38.finlit.online/predpriyatiya-ekonomika/metodyi-formirovaniya-tsenyi-53117.html