### Table 1: Results for the ring networks

"... In PAGE 7: ... A larger popula- tion and smaller selection pressure requires longer to converge but eventually does better by exploring more of the space. Table1 provides a summary of the results for the dif- ferent ring networks. Each row contains the results from one network.... In PAGE 8: ...Table 1: Results for the ring networks We report A, B and C for each network in columns 8-10 of Table1 . Figure 8 shows an example graph of this curve for the 16/8 ring network.... In PAGE 8: ... To nd a solution that is a fraction f of the way to the optimal solution therefore requires roughly (1 f) 1 C evaluations. In columns 11 and 12 of Table1 , we re- port the number of evaluations required to achieve 95% and 99% of the optimal solution, given by E95 = 20 1 C ; E99 = 100 1 C (3) Figure 8 shows these values for the 16/8 ring network. Random Networks - We next examine optimization performance on a set of randomly generated networks.... ..."

### Table 2 gives an example to computing the oblivious ratio.

"... In PAGE 12: ... Table2 : Oblivious Ratio, An Example 6.4 Tra c-oblivious Routing The optimal oblivious routing can be obtained by solving an LP with a polynomial number of variables, but in nitely many constraints.... ..."

### TABLE IV EFFICIENCY OF THE EXPLORATION WITH 5 ROBOTS Shape Of

2006

### Table 1: Results of all exploration techniques on the #0Cnite robot navigation task. #28a#29 In the middle column, the average

"... In PAGE 15: ... This technique, known as experiencereplay #5B26,27#5D, has been shown to be extremely powerful in aligning utility estimates through the U-table. Results are summarized in Table1 and Figure 6b. Parameters were carefully optimized by hand for all exploration techniques.... In PAGE 16: ... These costs are taken from and directly correspond to the experiments depicted in Figure 6b. #0F Search: Table1 a demonstrates the signi#0Ccant di#0Berence between undirected and directed exploration in terms of average time required for #0Cnding the goal #28#0Crst trial#29. Undirected techniques, all performing random exploration in the #0Crst trial, spent signi#0Ccantly more time until the goal was found.... In PAGE 16: ... Note that there is a signi#0Ccant gap between undirected #28#5Cb quot;, #5Cs quot;#29 and directed #28#5Cc quot;, #5Cr quot;#29 exploration, which also exists for all other directed techniques not plotted here. In all experiments we performed, directed exploration yielded much faster learning and always succeeded in #0Cnding close-to- optimal paths, whereas undirected exploration either mostly failed in identifying an optimal solution, or, as in the case of random exploration #28#5Cu quot;#29, succeeded only with tremendous exploration costs, #28 Table1 b#29. 6.... ..."

### Table 6.3: Exploration of a ring of XOR gates

1999

Cited by 7

### Table 6.3: Exploration of a ring of XOR gates

1999

Cited by 7

### Table 30. Large-Scale Planetary Exploration Robot Model Parameters

2003

"... In PAGE 13: ...able 29. Mission Performance in Small-Scale Planetary Exploration..................................................... 117 Table30 .... In PAGE 132: ...5 meters. The robot capabilities are summarized in Table30 by robot number. The experiment was run five times in order to evaluate the variability of the results.... ..."

Cited by 8

### TABLE II FINAL POSE ERRORS BY EXPLORATION POLICY FOR THE REAL ROBOT.

2003

Cited by 5

### Table 2. Allowables for both deterministic and probabilistic optimization

in AIAA-2002-1464 Probabilistic Design of a Plate-Like Wing to Meet Flutter and Strength Requirements

"... In PAGE 4: ... The first step was to carry out a deterministic opti- mization with margins on the allowables. The allow- ables for both the deterministic and probabilistic optimization are shown in Table2 . Standard methods were used to carry out the deterministic optimization.... In PAGE 7: ... Optimum design 2 is bounded by rising weight, by the flutter constraint, and by stress constraints from elements 6, 27, and 46. Probabilistic optimization The reliabilities of the two optimum deterministic designs were calculated using the probabilistic allow- ables given in Table2 . For designs 1 and 2, the reli- abilities are 0.... ..."

Cited by 1