Skip to main content

Featured

THE LEDGER Chapter 11 — A Message After All These Years

 THE LEDGER Chapter — A Message After All These Years It was the last day of the week. It was past 10 p.m. Aravind was sitting in his room. On the table were all the notes and reports from the investigations he had conducted throughout the week. After going through every detail of the cases he had investigated during the past week, Aravind prepared his Weekly Investigation Report. He carefully checked every piece of information. He went through the doubtful points once again. Then he completed the report. He placed it inside a file. He decided to send it by Registered Post from the post office the following day and kept the file safely on the table. He looked at the time again. It was already late. Aravind switched off the light. He went to bed. Exhausted from the week's investigations, he slowly fell asleep. --- The Next Morning As usual, Aravind woke up early. He completed his morning Surya Namaskar. After that, he went through his Kalaripayattu training. His entire body was cove...
Post a Comment
Read more

Forecasting Models – A Simple Explanation

 Forecasting Models – A Simple Explanation


Forecasting is an important part of Operations and Supply Chain Management. It helps a company estimate future demand and prepare its production, inventory, manpower and other resources.


What is Forecasting?


Forecasting means predicting future demand using past information.


For example, if a company studies its sales from the previous months, it can estimate how much customers may demand in the coming months.



---


1. Time Series Forecasting


Time Series Forecasting uses past demand data to estimate future demand.


Here, we mainly study how demand changes over time.


There are three important patterns:


Level


Trend


Seasonality




---


2. Level


Level means demand remains approximately at the same level over time.


For example, if monthly demand is around 1,000–1,100 units for several months, there is no major increase or decrease.


Methods such as Simple Average, Moving Average and Exponential Smoothing can be used for this type of data.



---


3. Trend


Trend means demand is continuously increasing or decreasing.


For example, a company's sales may increase from 1,000 units to 1,200, then 1,400 and then 1,600 units.


This indicates an increasing trend.


Linear Regression and Holt's Model can be used when the data shows a trend.



---


4. Seasonality


Seasonality means demand follows a regular pattern during particular periods.


For example:


Umbrella sales may increase during the rainy season.


Ice cream sales may increase during summer.


Many products have higher demand during festivals.



A seasonality index is used to understand how strongly a particular season affects demand.



---


5. Exponential Smoothing


Exponential Smoothing is a forecasting method that gives more importance to recent demand.


This is useful because recent demand may better represent the current market situation.


A small smoothing value gives relatively more importance to older information, while a larger value gives more importance to recent demand.



---


6. Holt's Model


Holt's Model is mainly used when the data has:


Level + Trend


It continuously updates the current level and the trend.


For example, if demand is steadily increasing, Holt's Model can use that trend to improve the forecast.



---


7. Winter's Model


Winter's Model is used when demand has:


Level + Trend + Seasonality


This makes it particularly useful for products whose demand changes according to seasons while also showing an increasing or decreasing trend.


One important advantage of Winter's Model is that it gives more importance to recent information through exponential smoothing.


Easy way to remember:


Holt's Model → Level + Trend


Winter's Model → Level + Trend + Seasonality



---


8. Causal Models


A causal model is used when one factor affects another factor.


For example:


Rainfall → Agricultural Production


Here, rainfall can influence agricultural production.


Other examples include:


Advertising → Sales


Price → Demand


Temperature → Electricity consumption



Causal models help companies understand how changes in one factor can influence demand.



---


9. Measuring Forecast Accuracy


After making a forecast, we need to check:


"How good is our forecast?"


Several measures can be used, including:


Mean Absolute Deviation (MAD)


Mean Squared Deviation (MSD)


Mean Percentage Deviation


Mean Absolute Percentage Error (MAPE)



Generally, a smaller forecasting error indicates a better forecast.



---


10. Choosing the Right Forecasting Model


The forecasting model should be selected according to the behavior of the data.


If the data has only Level:

Use Simple Average, Moving Average or Exponential Smoothing.


If the data has Level + Trend:

Use Linear Regression or Holt's Model.


If the data has Level + Trend + Seasonality:

Use a Seasonal Model or Winter's Model.


So, we should first understand the pattern in the data and then select the appropriate forecasting method.



---


11. From Forecasting to Production Planning


Forecasting tells us how much demand we may have in the future.


The next question is:


> How much should the company produce?


If you like to support my content use this link


This is where Aggregate Production Planning comes in.


The company needs to decide:


How much to produce during regular working hours


How much overtime production is required


How much inventory should be maintained


Whether any shortage may occur



The main objective is to meet customer demand while keeping total cost as low as possible.


The major costs include:


Regular production cost


Overtime cost


Inventory carrying cost


Shortage cost




---


Conclusion


Forecasting helps a company prepare for future demand. By studying level, trend and seasonality, the company can select an appropriate forecasting method.


Simple Average, Moving Average and Exponential Smoothing are useful for level data. Holt's Model is useful for level and trend. Winter's Model is useful when level, trend and seasonality are present.


The forecast can then be used for Aggregate Production Planning, helping the company decide how much to produce, how much inventory to maintain and how to minimize production-related costs.

Comments

Post a Comment