IE 375 - Production Planning Homework Assignment 1 - October 10 The past demand data for taxi rides at Bilkent Campus is given in the following table. Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Monday 18 20 20 21 23 25 Tuesday 12 14 15 15 17 16 Wednesday 14 15 16 18 18 19 18 19 20 21 22 Thursday 17 Friday 25 27 29 30 29 32 (a) (10 points) Use the moving averages method to forecast the demand in periods 11-30, taking into account the latest demand information. Find the value of N = {5, 7, 10} that minimizes the MAD calculated via the forecast errors in periods 11-30. (b) (10 points) Use the exponential smoothing method to forecast the demand in periods 11-30, taking into account the latest demand information. Use the demand observations in periods 1-10 to initialize F1. Find the value of a {0.1, 0.2, 0.3} that minimizes the MAD calculated via the forecast errors in periods 11-30. (c) (20 points) Use Holt's method to forecast the demand in periods 11-30, taking into account the latest demand information. Use the demand observations in periods 1-10 to initialize S10 and G0. Assume = 0.1. Find the value of a {0.1,0.2, 0.3} that minimizes the MAD calculated via the forecast errors in periods 11-30. (d) (20 points) Use Winter's method to forecast the demand in periods 11-30, taking into account the latest demand information. Use Winter's initialization procedure with the demand observations in periods 1-10 to initialize S10, G10, C6, C7, C8, C9, and C10. Assume 30.1. Find the values of a (0.1,0.2, 0.3} and {0.1, 0.2} that minimize the MAD calculated via the forecast errors in periods 11-30. (e) (20 points) Answer the question in part (d) with a different initialization procedure: First deseasonalize the series from period 1 to 10 and then fit a simple linear regression. (f) (20 points) Which of the above methods yields the least MAD? Forecast the total demand in week 7 with the best method and best parameter values you found.