Question

Siberian Charter Adventures is an independent Canadian company focused on providing unique charter tours to Siberia. The company has averaged between 1,000 and 1,500 tourists per month, and the operations manager wants to develop a forecast model to aid in planning. The operations manager has provided you with historical tourist data for the 24 months in 2014 and 2015.

1. Download the accompanying data file (MGSC369_R6_A4_datafile) and prepare a set of forecasts for each of the following three methods. Please read the instructions in the Excel^TM file carefully before you begin. Note that the forecasts can be created using Microsoft Excel, and you should submit your spreadsheet with your assignment for grading

Month	Actual Demand (# of Tourists)
Jan-14	1344
Feb-14	1339
Mar-14	1415
Apr-14	1213
May-14	1412
Jun-14	1318
Jul-14	1018
Aug-14	1250
Sep-14	1176
Oct-14	1062
Nov-14	1137
Dec-14	1391
Jan-15	1192
Feb-15	1425
Mar-15	1320
Apr-15	1256
May-15	1288
Jun-15	1420
Jul-15	1478
Aug-15	1198
Sep-15	1203
Oct-15	1002
Nov-15	1246
Dec-15	1339

a) Three-month moving average

b) Exponential smoothing based on what you believe to be an appropriate alpha constant. Provide a brief justification

c) Trend-adjusted exponential smoothing based on what you believe to be appropriate alpha and beta constants. Provide a brief justification for your choices.

2. Complete the following error summary table of the three forecasts, and then comment on the results. Identify which forecast method you believe the company should adopt.

	MAD	MSE	MAPE
three-month moving average
Exponential smoothing
Trend-adjusted exponential smoothing

Answer 1

	MAD	MSE	MAPE
MOVING AVERAGE	79.22146636	22811.31217	9.8093%
EXPONENTIAL SMOOTHING	109.67695	19361.88563	9.0556%
TREND ADJUSTED SMOOTHING	111.568022	21258.756	9.1412%

For exponential smoothing take alpha=0.2761 as the actual values are varying greatly.

For Trend Adjusted exponential smoothing take alpha=0.2761 and beta=0.1 .Here we can notice that MAPE from trend adjusting is more than exponential smoothing and hence to minimize the MAPE error take beta=0.

Initially beta=0.1 was taken to give weight to the trend.

THREE MONTH MOVING AVERAGE

EXPONENTIAL SMOOTHING WITH APLHA=0.27

EXPONENTIAL TREND SMOOTHING WITH APLHA=0.27 and beta=0.1