### Table 4 Estimated expected total number of observations for scheduling data.

"... In PAGE 22: ...0. Table4 summarizes the results along with the number of hypothe- ses and the relative difficulty (o/ amp;) of each data set. The principle result is that STOP1 and STOP2 substantially exceeded the performance of the other algo- rithms except on one case.... In PAGE 28: ...0) 2347 EL2(0.5) 2211 Table4 : Direct Comparison of all four algorithms These results show that the algorithms produce roughly comparable utilities, the difference in utilities is smaller than the smallest indifference interval specified to the interval-based algorithms. From this comparison we must conclude that, at least in the case of this NASA scheduling application, there is little difference between the interval-based and expected loss approaches, both in terms of ex- pected improvement and in terms of sample complexity.... ..."

### Table 1 Estimated expected total number of observations in the least favorable configuration.

### Table 6 Estimated expected total number of observations and expected loss of an incorrect selection

### Table 4: Estimated expected total number of observations to rank DS-2 spacecraft designs. Achieved probability of correct ranking is shown in parenthesis.

1999

"... In PAGE 15: ...with those cases in which the penetrator does not penetrate being assigned zero utility. As shown in Table4 , both PAC algorithms significantly outperformed the Turnbull algorithm, which is to be expected because the hypotheses are somewhat correlated (via impact orien- tations and soil densities). Table 5 shows that the EL,,, expected loss algorithm effectively bounded actual loss but the EL,,ii algorithm was irlcorlsist,ent.... ..."

Cited by 2

### Table 2: Estimated expected total number of observations of EL algorithms in stepped means configuration. Observed average loss of produced rankings.

1999

Cited by 2

### Table 2.4: Expected total number of customers in Example 1. Case g = 1 g = 2 g = 3 g = 4 g = 5 g = 10

2003

### Table 2.5: Expected total number of customers in Example 2. Case g = 1 g = 2 g = 3 g = 4 g = 5 g = 10

2003