[sahana-maindev] the selected Algorithm to Optimize the Inventory

[mahesh <ma...@op...>]

Hi All,

To make predictions of the available stock amount in warehouses I have
found an algorithm known as "Single Exponential Smoothing". 

The Exponential Smoothing is the most used forecasting method and It is
a weighted average method that gives more weight to the recent data. It
forecasts (f) for a period (t) using a variable (alpha α) as a known
smoothing constant. Alpha is a number between zero and one. The formula
is: 

"f(t+1)=(1-α)ft+α*at"


The variable a(t) stands for the actual data from period t. This formula
is used to predict data one period ahead. 

f(t+1)= forecast value at period (t+1)

ft= forecast value in period t

at= actual value at period t

(Alpha) = the smoothing constant

If Alpha is set to 1, the forecast for the next period is based entirely
on the actual value from the last period. If Alpha is set to 0, the
actual value from the last period is completely ignored. 



	With the following section I have discussed why the Single Exponential
Smoothing algorithm was selected to predict the stock outage in
warehouses in disaster management situations. 

1.The Exponential Smoothing algorithm is used to predict data one 
  period ahead. In disaster like situations the 
  warehouses normally not going to store items for very long period. 
  There fore predicting the stock-outage for a long period ahead is not 
  necessary. In such a case the single Exponential 
  Smoothing algorithm can be applied to predict the stock outage in 
  warehouses. 

2.In Disaster like situations probably donations come soon after the
  disaster. There fore stock amount is rapidly increased soon after  
  the disaster. But some times after the disaster the amount of  
  donations are decreased and therefore stock
  amount would come lower. The prediction process of the algorithm is
  based on most recently taken data. The advantage of that is when a  
  new  observation becomes available, a new prediction can be 
  done  by dropping the oldest observation and including the newest one.

3.The formula f(t+1)=(1-α)f(t)+α*a(t) is used to predict data one  
  period  ahead. For example if we have the data for time period "t" 
  then we can make prediction for time period "t+1","t+2",.....,"t+n" 
  respectively. Therefore when make predictions for the data for long  
  time period ("t+n") there would be more and complex calculations. 
  There fore it takes much more processing power. So the algorithm is 
  useful to make predictions for a short time period.  

These factors helped me to select the Exponential Smoothing algorithm to
predict the stock outage in warehouses.

Thank You,
Regards,
Mahesh.