Problem 1: Fashion Retail - "TrendWear" (Clothing & Accessories)

"TrendWear," a fast-fashion retailer with multiple outlets in Mumbai, needs to forecast monthly
demand for its best-selling "UrbanFit" jeans. The sales data (in units) for the last 12 months is:

Month Sales (Units)

1 3200
2 3500
3 3700
4 4000
5 3100
6 3400
7 2900
8 3300
9 3800
10 3200
11 3000
12 3100

Forecast the demand for "UrbanFit" jeans for the next 3 months using a 3-month moving
average and simple exponential smoothing with an appropriate smoothing factor. Evaluate and
compare the MAD, MAPE, MSE, bias, and tracking signal (TS). As an Inventory Manager at
"TrendWear," which forecasting method would you recommend for inventory planning, and
why?

Problem 2: Electronics Manufacturing - "BrightTech" (LED Televisions)

"BrightTech," an electronics manufacturer producing LED televisions, wants to predict monthly

production requirements for its 40-inch Smart LED TV. The monthly order data (in units) for
the past 10 months is:

Month Orders (Units)

1 5000
2 5300
3 5800
4 6200
5 4500
6 4900
Month Orders (Units)
7 4200
8 4600
9 5500
10 4800

Estimate the demand for the next 3 months using a 3-month moving average and simple
exponential smoothing. Compute MAD, MAPE, MSE, bias, and TS to assess accuracy. As a
Production Planning Manager at "BrightTech," which method would you choose for production
scheduling, and why?

Problem 3: Ride-Hailing Services - "GoCab" (Urban Mobility)

"GoCab," a ride-hailing service in Bengaluru, wants to predict the number of daily ride bookings
to optimize driver availability. The daily ride data for the last 15 days is:

Day Rides (Bookings)

1 12000
2 12500
3 13500
4 15000
5 11000
6 12200
7 10500
8 11500
9 14000
10 11800
11 10800
12 11200
13 12500
14 11000
15 11800

Predict the ride demand for the next 5 days using a 3-day moving average and simple
exponential smoothing. Compute MAD, MAPE, MSE, bias, and TS. If you were the
Operations Manager at GoCab, which method would you recommend for driver allocation and
why?
Problem 4: Dairy Industry - "PureMilk" (Fresh Milk Distribution)

"PureMilk," a dairy company distributing fresh milk across Chennai, needs to forecast weekly
demand for its 500ml milk packets to manage procurement and distribution efficiently. The
weekly sales data (in liters) for the last 12 weeks is:

Week Sales (Liters)

1 15000
2 16200
3 17500
4 18500
5 14000
6 15500
7 13000
8 14800
9 17000
10 14500
11 13500
12 14000

Estimate the demand for the next 4 weeks using 4-week moving average and simple
exponential smoothing. Compare MAD, MAPE, MSE, bias, and TS. As a Distribution
Manager at PureMilk, which forecasting technique would you suggest for managing supply
chain operations, and why?
Problem 1: Retail Inventory Forecasting – “TrendyCloth” (Clothing Retailer)

"TrendyCloth," a clothing retailer, wants to forecast weekly demand for their best-selling
"Casual Shirts" to optimize stock levels. The past 10 weeks' sales data (in units) is:

Week Sales (Actual Demand)

1 500
2 550
3 600
4 620
5 480
6 530
7 450
8 490
9 580
10 520

Tasks:

1. Forecast demand for the next 3 weeks using 3-week Moving Average and Simple
Exponential Smoothing (α = 0.2).
2. Compute Mean Absolute Deviation (MAD), Mean Squared Error (MSE), Mean
Absolute Percentage Error (MAPE), Bias, and Tracking Signal (TS).

3-Week Moving Average Forecast Calculation

Formula:

Moving Average Forecast = (Sum of last 3 weeks' demand) / 3

Sales Forecast
3-Week Moving Absolute Error Percentage
Week (Actual Error (Actual
Average Forecast Error Squared Error (%)
Demand) - Forecast)
1 500 - - - - -
2 550 - - - - -
3 600 - - - - -
(500+550+600)/3 =
4 620 620 - 550 = 70 70 4900 11.29
550
(550+600+620)/3 = 480 - 590 = -
5 480 110 12100 22.92
590 110
 Sales Forecast
3-Week Moving Absolute Error Percentage
Week (Actual Error (Actual
Average Forecast Error Squared Error (%)
Demand) - Forecast)
(600+620+480)/3 = 530 - 566.67 =
6 530 36.67 1346.67 6.92
566.67 -36.67
(620+480+530)/3 = 450 - 543.33 =
7 450 93.33 8710.89 20.74
543.33 -93.33
(480+530+450)/3 = 490 - 486.67 =
8 490 3.33 11.11 0.68
486.67 3.33
(530+450+490)/3 =
9 580 580 - 490 = 90 90 8100 15.52
490
(450+490+580)/3 = 520 - 506.67 =
10 520 13.33 177.78 2.56
506.67 13.33
(∑ Errors) = -
Sum - - 416.33 35346.45 80.63%
63.67

Exponential Smoothing Forecast Calculation (α = 0.2)

Formula:

Forecast = α (Actual Demand) + (1 - α) (Previous Forecast)

Assumption: Initial forecast for Week 1 = 500 (same as actual demand).

Sales Forecast Error

Exponential Smoothing Absolute Error Percentage
Week (Actual (Actual -
Forecast Error Squared Error (%)
Demand) Forecast)
1 500 500.00 - - - -
2 550 0.2(550) + 0.8(500) = 510 550 - 510 = 40 40 1600 7.27
3 600 0.2(600) + 0.8(510) = 528 600 - 528 = 72 72 5184 12.00
0.2(620) + 0.8(528) = 620 - 548.4 =
4 620 71.6 5126.56 11.55
548.4 71.6
0.2(480) + 0.8(548.4) = 480 - 535.52 = -
5 480 55.52 3082.82 11.57
535.52 55.52
0.2(530) + 0.8(535.52) = 530 - 534.42 = -
6 530 4.42 19.54 0.83
534.42 4.42
0.2(450) + 0.8(534.42) = 450 - 517.54 = -
7 450 67.54 4561.08 15.01
517.54 67.54
0.2(490) + 0.8(517.54) = 490 - 511.03 = -
8 490 21.03 442.26 4.29
511.03 21.03
 Sales Forecast Error
Exponential Smoothing Absolute Error Percentage
Week (Actual (Actual -
Forecast Error Squared Error (%)
Demand) Forecast)
0.2(580) + 0.8(511.03) = 580 - 523.82 =
9 580 56.18 3156.17 9.69
523.82 56.18
0.2(520) + 0.8(523.82) = 520 - 523.06 = -
10 520 3.06 9.36 0.59
523.06 3.06
(∑ Errors) = -
Sum - - 391.35 23583.79 72.81%
67.61

Error Metrics Comparison

Calculations:

• MAD (Mean Absolute Deviation) = Sum of Absolute Errors / n

• MSE (Mean Squared Error) = Sum of Squared Errors / n
• MAPE (Mean Absolute Percentage Error) = (Sum of Percentage Errors / n) * 100
• Bias = Sum of Errors / n
• Tracking Signal (TS) = Bias / MAD

Method MAD MSE MAPE Bias TS

Moving Average 41.63 3534.65 8.06% -6.37 -0.15
Exponential Smoothing 39.14 2358.38 7.28% -6.76 -0.17

Recommendation

Exponential Smoothing performs better as it has lower MSE, MAPE, and MAD.
It adapts better to fluctuations compared to the Moving Average method.
Recommended for inventory management at “TrendyCloth.”