Implementation of the Moving Average Method for Forecasting Inventory in Cv. Tre Jaya Perkasa

The supply chain is an


Introduction
The supply chain is an organization's place to distribute production goods and services to customers.This chain is a network of various organizations that are interrelated and have the same goal, namely to organize the procurement or supply of goods.Entities that play a role in the supply chain are suppliers, manufacturers, distributors, retail outlets, and customers.[1] [2] example, the case of the NIKE company, which recently experienced problems in the supply chain due to production and distribution delays, resulting in increased sales and administrative costs [5].
Forecasting is an attempt to predict future conditions by studying data sets from the past [6] [7].The use of forecasting in trading companies can help companies to find out the number of requests that will come, so that companies can consider whether to add inventory or not [8].
CV. Tre Jaya Perkasa is a company engaged in the distribution of goods.CV.Tre Jaya Perkasa already has more than 1000 customers who pick up goods for resale.The number of sales transactions will affect the inventory of goods in the warehouse.Because many transactions are carried out, inventory control is needed.Inventory control is very important for distributor companies.This is useful for minimizing storage costs that must be borne by the company, and meeting customer demands [9].However, because demand is difficult to know with certainty, stock reserves become an alternative to meet more changes in demand.Market or customer demand that cannot be fulfilled will result in out of stock on certain products.With this problem, we need a system that can predict the number of items that must be ordered from suppliers, so that there will be no shortage or excess stock of goods in the warehouse.
One method in forecasting that can be used is the Moving Average method.The Moving Average method uses time as the basis for forecasting.Past historical data is the basic material used to be described in forecasting.Furthermore, to calculate the level of forecasting error is calculated by MAD (Mean Absolute Deviation), MSE (Mean Squared Error), and MAPE (Mean Present Error) [3] [10].
Similar research has been carried out by Hayuningtyas, in this study combining the weighted moving average method and the double exponential smoothing method, the data used for this research is sales data for one year 2016.In this study, the comparison of forecasting results using both methods is described.inventory for the next period 52 with Weighted Moving Average and 60 with Double Exponential Smoothing.Both of these methods have a Mean Square Error value.Where the Weighted Moving Average error value is 0.114 and the Mean Square Error error value is 6.12, the smallest error value [9].
While the research conducted by Rizal Rahman, in this study compared 2 forecasting methods, namely the moving average method and the Exponential Smoothing method, the data used is consumer demand data in 2017.In this study, a comparison between the two methods is described, the results of demand forecasting for January 2018 using the moving average method is 76.999.67 pcs, while the exponential smoothing method = 0.9 is 78.146 pcs.Between the two methods, the method that has a smaller forecast error value is the exponential smoothing method =0.9.[11]

Research Method
The primary data collection methods used in this research are: • Observation In this method the researchers made a direct visit to CV. Tre Jaya Perkasa and conducted data collection from existing processes in the field to obtain the data needed in this research.• Interview In this method, a question and answer process is carried out with several sources, to obtain the data used in this study.• Study of literature Studying and understanding books, journals and others related to this research.
The secondary data from this research is the transaction history data collection of Cv.Tre Jaya Perkasa from April 2020 to June 2021.For this study, the data used is sales data for the TRICK POTATO BISCUIT BBQ 24 BOX X 10X18 product.Sales data from these products will be tested for the 3 period and 5 period moving average method.Sales data of the TRICK POTATO BISCUIT BBQ 24 BOX X 10X18, can be seen in Table 1.The method used to predict the amount of stock using the type of quantitative forecasting, namely the moving average method Forecasting Approach This study uses a quantitative approach, namely research that emphasizes analysis on numerical data that is processed by statistical methods to be able to predict future conditions or trends.Quantitative research allows for generalization to the results, which are calculated by statistical analysis [12] [13].
The quantitative approach can be separated into 2 criteria, namely the Time Series approach, which is a design stage that makes a process to obtain accurate planning results as a unit that cannot be found and no further process is needed to find it.Then the second is the causal explanatory technique, which is a method in which it considers the forecasting results to have a causal relationship by utilizing system inputs [14].

B. Determining the Forecasting Model
To determine the forecasting model to be used, it is necessary to conduct data analysis.The historical data used is stationary or non-stationary data.Stationary data is historical data that has an average and has a tendency to move towards the average.Stationary data occurs when there is no sharp increase or decrease in the data [15].Non-stationary data is historical data that has a trend, tends to move up or down.
In Figure 1, you can see a graph of sales of the TRICK POTATO BISCUIT BBQ 24 BOX X 10X18, on the graph you can see how the data pattern is to determine the forecasting model that can be used.It can be seen in Figure I that the data does not experience a significant trend, sales data tends to move towards the average.Because the data is not trend data, so for the forecasting process it can be used using the moving average method.

C. Moving Average
Moving Average is a forecasting method that is carried out by using a group of data that has a value, then looking for the average of that group of values for forecasts in the future period [3].
Forecasting with this moving average method assumes that all values in historical data have the same weight value in calculating the forecast value in the next period.Forecasting using the moving average method is the simplest forecasting, because it does not use weighting in forecasting calculations.Although simple, this method is quite effective for knowing market trends [16] The characteristics of the Moving Average Method are as follows: • Historical data over a certain period of time is needed to determine forecasts for the future period.For example with a 3 month moving average, to forecast the 4th month.• The longer the moving average time, the better the moving average will be.
The mathematical equation of the Moving Average method is: Description:   : Moving Average for period t  +1 : Forecast for period t+1   : Real value period t  : the number of limits in the moving average

D. Measuring Forecasting Error
The measure of the accuracy of the forecasting results is a measure of the degree of difference between the forecasting results and the actual value.This measure of forecasting accuracy can be measured by [17]: 1) MAD (Mean Absolute Deviation) Mean Absolute Deviation is used to measure the accuracy of the forecast by averaging the magnitude of the forecast error where each forecasting has an absolute value for each error [18].The formula for calculating MAD is as follows: MSE is the second way to measure forecasting error.MSE is the average of the squared differences between the actual values and the forecast values [19].The formula for calculating MSE is as follows: 3) MAPE (Mean Absolute Percent Error) MAPE is the average of the absolute difference (absolute) between the forecast and actual values, then displayed as a percentage of the actual values [20].The formula for calculating MAPE is:

Results and Discussion
To forecast the inventory of goods in the future period, namely the period of July 2021, the data used is sales data from April 2020 to June 2021 from the products most frequently purchased by customers, namely TRICK BISKUIT KENTANG BBQ 24 BOX X 10X18 products.Forecasting will be done using the 3 period and 5 period moving average method.1) Forecasting with 3 period The results of sales forecasting for the July 2021 period will be a reference for determining the amount of inventory that must be provided by the company for transaction processing in July 2021.The sales forecasting process in July can be seen in Table 2.The results of forecasting sales of TRICK BISKUIT KENTANG BBQ 24 BOX X 10X18 products in July 2021 using a 5-period moving average, can be seen in Table 3.

Figure 1 .
Figure 1.Product Sales Chart Trick Potato Biscuit Bbq 24 Box X 10x18

Table 2 .
Forecasting Results Using 3 PeriodsFrom the table above, it is known that the forecasting results in July with 3 periods are 1.844.After doing the forecasting, the next step is to calculate MAD, MSE and MAPE.

Table 3 .
Calculation Results of Mad, Mse, And Mape

Table 4 .
Forecasting Results Using 5 PeriodsFrom the table above, it is known that the forecasting results in July with 5 periods are 1.970.After doing the forecasting, the next step is to calculate MAD, MSE and MAPE.

Table 5 .
Calculation Results of MAD, MSE, and MAPE