### Table 2-1. Execution Times for Dynaplan Automatic Route Planner

1991

"... In PAGE 37: ...Table2 -2 summarizestheTF/TA/NOE/FCSrequirementsobtained from the literature search. It is expected that, for reasons mentioned above, the processor count indicated in the table provides a throughput that significantly exceeds that needed by the application.... ..."

Cited by 1

### Table 1: Overall success rate per planner and per environment

2004

"... In PAGE 7: ... We postulate, therefore, that for increasingly complex environments the RBF planner will eventually outperform the Neighbourhood planner. Table1 summarises the success rates of the three planners in the two environments, show- ing the significant difference between the original planner and the two enhanced planners. Table 1: Overall success rate per planner and per environment... ..."

Cited by 1

### Table 1: Overall success rate per planner and per environment

2004

"... In PAGE 7: ... We postulate, therefore, that for increasingly complex environments the RBF planner will eventually outperform the Neighbourhood planner. Table1 summarises the success rates of the three planners in the two environments, show- ing the significant difference between the original planner and the two enhanced planners. Table 1: Overall success rate per planner and per environment... ..."

Cited by 1

### Table 2: The percentage of successful connections (by different local planners).

1998

"... In PAGE 10: ...lanners) in each version. In each case, the same number of connections were attempted. The percentage of connections made can be seen as a rough approximation of . As can be seen in Table2 , for each environ- ment, and for each local planner, the total number of successful connections decreases as the environment gets harder. Remember, the hardness we define is not necessarily related to the connectivity of the resulting roadmap.... ..."

Cited by 59

### Table 2: Success rate of the automatic vision system.

### Table 5: Time-to-success (left side) across the non-dominated planners; the planners are ordered according to selection by the Greedy-Set-Cover. The right side shows the percentiles for the planners.

"... In PAGE 4: ... We propose using the RTDs to arrive at a more informed allocation strategy that will make good use of extra time when justi ed. Table5 (left) shows the log distributions of time-to-success for the non- dominated planners. It is clear that we do indeed see heavy tails in most of the planner RTDs.... In PAGE 5: ... A simple observation leads to a remarkably successful al- location strategy that does not use either of the above anal- yses or models. As shown in Table5 (right), all but one achieve the 80th percentile at 10 seconds. At 100 seconds, one planner each achieves the 89th, 94th, and 97th per- centiles, ve reach the 98th percentile, and three achieve the 99th percentile.... ..."

### Table 6. Comparison of Centralized and Decoupled Planners

"... In PAGE 18: ... The de- coupled planner with pairwise coordination had up to 50,000 milestones to search each of the two-dimensional coordina- tion space. Table6 lists the number of failures and the average run- ning times of the two decoupled planners over 20 runs on each of problems I, II, and III, with 2, 4, and 6 robots. (The average running times are computed only over the successful runs.... ..."

### Table 10: Max, mean, median and standard deviations (Sd) for the computation times to success and failure for each planner.

2002

"... In PAGE 18: ... Across the planners and the problem set, we found that the distributions were skewed (approximately log normal with long right tails) and that the planners were quick to declare success or failure, if they were going to do so. Table10 shows the max, mean, median and standard deviation for success and failure times for each of the planners. The di erences between mean and median indicate the distribution skew, as do the low standard deviations relative to the observed max times.... In PAGE 22: ... Not surprisingly, faster processor and more memory nearly always lead to better performance. Somewhat surprisingly, the di erence is far less than the doubling that might be expected; the mean di erences are much less than the mean times on the faster processor (see Table10 for the mean solution times). Also, the e ect seems to vary between the planners.... ..."

Cited by 10

### Table 10: Max, mean, median and standard deviations (Sd) for the computation times to success and failure for each planner.

2002

"... In PAGE 18: ... Across the planners and the problem set, we found that the distributions were skewed (approximately log normal with long right tails) and that the planners were quick to declare success or failure, if they were going to do so. Table10 shows the max, mean, median and standard deviation for success and failure times for each of the planners. The di erences between mean and median indicate the distribution skew, as do the low standard deviations relative to the observed max times.... In PAGE 22: ... Not surprisingly, faster processor and more memory nearly always lead to better performance. Somewhat surprisingly, the di erence is far less than the doubling that might be expected; the mean di erences are much less than the mean times on the faster processor (see Table10 for the mean solution times). Also, the e ect seems to vary between the planners.... ..."

Cited by 10

### Table 10: Max, mean, median and standard deviations #28Sd#29 for the computation times to success and failure for each planner.

2002

"... In PAGE 18: ... Across the planners and the problem set, we found that the distributions were skewed #28approximately log normal with long right tails#29 and that the planners were quicktodeclare success or failure, if they were going to do so. Table10 shows the max, mean, median and standard deviation for success and failure times for each of the planners. The di#0Berences between mean and median indicate the distribution skew, as do the low standard deviations relative to the observed max times.... In PAGE 22: ... Not surprisingly, faster processor and more memory nearly always lead to better performance. Somewhat surprisingly, the di#0Berence is far less than the doubling that might be expected; the mean di#0Berences are much less than the mean times on the faster processor #28see Table10 for the mean solution times#29. Also, the e#0Bect seems to vary between the planners.... ..."

Cited by 10