Logical Decisions - Classweb
Logical Decisions - Classweb Logical Decisions - Classweb
percent show destructive interaction, while Ps greater than 50 percent show constructive interaction. Since alternative B is preferred when P is equal to 50 percent you must make alternative B less desirable. You do this by reducing P, so that the probability of getting the least preferred level on both measures is increased. Since your P will be less than 50 percent, your answer will show destructive interaction between the Power and Fuel Economy measures. To find a particular level for P you can try different probabilities until you zero in on one that makes alternatives A and B seem equally desirable. If you try P = 5 percent, you choose alternative A because the chance of getting your preferred option in alternative B is too remote. You might feel the same way even if P = 30 percent. At 40 percent, alternative B still seems preferable to alternative A. If this is the case, then you might select P = 35 percent as the level of P that makes it most difficult to choose between alternatives A and B. 7. Answer the probability question by replacing the default of 50 percent with 35 percent. 8. Click “Equal.” LDW will use this answer when computing the MUF for the Performance goal. LDW computes the utility of the Performance goal using the multiplicative MUF formula described on page 9-55. You can see the new scaling constants for performance by selecting the Review::Weights::Scaling Constants option. In the table, you will see that the big K for performance is 2.45 and that the scaling constants for the two members are both 0.35. Since the scaling constant big K is greater than zero, destructive interaction is indicated. This is also shown by the fact that the individual weights (small ks) for the members of the Performance goal (the Power and Fuel Economy measures) do not sum to one as they do for the members of the Best Truck goal, which has neutral interaction. If we now use the Results::Rank Alternatives option to see the ranking of the alternatives on the "Buy the Best Truck" goal, we 5-18 Section 5 -- Advanced Tutorial
see that the rankings for all the alternatives have gone down slightly. The "Coyote's" ranking goes from 0.681 to 0.647, the "Mountain Lion's" ranking goes from 0.592 to 0.562, and the "Wolf's" ranking goes from 0.376 to 0.310. Thus the addition of interactions seems to have had a somewhat larger negative effect on the "Wolf" alternative than on the others. This ends the Advanced Tutorial. Section 5 -- Advanced Tutorial 5-19
- Page 27: S E C T I O N Basic Tutorial 4
- Page 30 and 31: Figure 4-1. Tutorial overview. 4-2
- Page 32 and 33: Now lets make sure the alternatives
- Page 34 and 35: Assume you have decided that you wi
- Page 36 and 37: Next we will enter the measures for
- Page 38 and 39: Defining Preferences The alternativ
- Page 40 and 41: 1. Select the Assess::Common Units
- Page 42 and 43: almost equally unacceptable, while
- Page 44 and 45: 1. Select the Assess::Common Units
- Page 46 and 47: When you do this, the tradeoff grap
- Page 48 and 49: 1. Select "Performance" and "Price"
- Page 50 and 51: 8. Click on the "Equal" button to t
- Page 52 and 53: Figure 4-11. Display generated by R
- Page 54 and 55: Figure 4-13. Overall ranking for tr
- Page 56 and 57: Figure 4-15. Graph showing sensitiv
- Page 58 and 59: You can see the completed introduct
- Page 61 and 62: Advanced Tutorial This tutorial sec
- Page 63 and 64: Probabilities There is a problem wi
- Page 65 and 66: This information indicates that the
- Page 67 and 68: A uniform distribution is defined b
- Page 69 and 70: On the left is a list of the possib
- Page 71 and 72: Figure 5-2. Example of Results::Unc
- Page 73 and 74: screen a large database for the alt
- Page 75 and 76: 9. LDW will ask if you want to appe
- Page 77: commit to buying their truck before
- Page 81: S E C T I O N Using LDW 1: Structur
- Page 84 and 85: The Edit::Insert option. The Insert
- Page 86 and 87: ! Summary -- view a dialog box that
- Page 88 and 89: structure like an organization char
- Page 90 and 91: If you check the Show Assessment St
- Page 92 and 93: saves it for later pasting. When yo
- Page 94 and 95: You can create a new Matrix view by
- Page 96 and 97: The quick entry view shows the alte
- Page 98 and 99: Structuring Goals The goals in an L
- Page 100 and 101: these two fields to describe each m
- Page 102 and 103: the hierarchy. In the Matrix view,
- Page 104 and 105: ! Point Estimate -- use a single nu
- Page 106 and 107: Figure 6-12. Example of a measure l
- Page 108 and 109: Figure 6-14. Example of a measure l
- Page 110 and 111: pass, LDW replaces each probabilist
- Page 112 and 113: Figure 6-17. Dialog box for definin
- Page 114 and 115: Figure 6-18. Measure Category Dialo
- Page 117: S E C T I O N Using LDW 2: Assessin
- Page 120 and 121: measure utility functions for the g
- Page 122 and 123: information as possible when you ch
- Page 124 and 125: measures with categories require mu
- Page 126 and 127: The "Reset" button deletes any asse
see that the rankings for all the alternatives have gone down<br />
slightly.<br />
The "Coyote's" ranking goes from 0.681 to 0.647, the "Mountain<br />
Lion's" ranking goes from 0.592 to 0.562, and the "Wolf's" ranking<br />
goes from 0.376 to 0.310. Thus the addition of interactions seems<br />
to have had a somewhat larger negative effect on the "Wolf"<br />
alternative than on the others.<br />
This ends the Advanced Tutorial.<br />
Section 5 -- Advanced Tutorial 5-19