### Table 9: Nodes in branch and bound tree

### Table 7 - Results obtained with the branch and bound algorithm

"... In PAGE 16: ...) (pc_sp) and of sequences _r (pc_r) examined by the enumeration process are shown in Table 7. lt;Insert Table7... ..."

### Table 4: Statistical informations Branch amp; Bound

1995

### Table 9: Nodes in branch and bound tree

2000

### Table 4: Results for branch-and-bound algorithm. Type Date Opt Branch amp;Bound

"... In PAGE 24: ... The third column contains the value of the integer optimum. The last four columns show the compu- tation times (in seconds), the number of branching nodes needed to solve the problems, the percentage of problem instances that are solved in less than one second, and the percentage of instances that are solved in more than 10 seconds (Table 3 contains these values for the branch-and-price al- gorithm; Table4 for the branch-and-bound algorithm). All values in these tables are average values over the 1081 pairs of ASs, so they contain results of over 20,000 problem instances.... In PAGE 24: ... While we could run the branch-and-price algorithm to completion on all pairs in all graphs, we had to terminate the branch-and-bound algorithm on a few pairs (at most 10 out of 1081 pairs in each of the graphs) after several hours of computation time. The running- time and the number of branching nodes shown for the branch-and-bound algorithm in Table4 are thus the averages over the pairs for which the algorithm could be run to completion. From Tables 3 and 4 we conclude that, on the average, both algorithms perform well on the selected pairs of ASs.... ..."

### Table 1: Computational results for the basic case mcf branch amp;bound branch amp;cut lp branch amp;bound

2007

"... In PAGE 33: ... All computations were done on a Pentium IV 2 GHz computer, with 512 Mb RAM. In Table1 , the results for the basic TQD problem are presented. Each al- gorithm solves all instances in a reasonable amount of time; random instances seem to be harder to solve than the structured ones.... ..."

Cited by 2

### Table 1: Upper bound for the required number of GC increments per block allocated as a function of the upper bound of the reachable memory.

"... In PAGE 4: ... This guarantees that a cycle finishes at the latest after the number of increments performed exceeds the amount of allocated memory, as is required by this approach (see [22] for the details). Table1 illustrates the worst- case number of increments that are required for the allocation of one block, as a function of the upper bound of reachable memory k = K / M. On the PowerPC processor, a marking or sweeping of one block takes about winc = 80 (compiled) machine instructions.... ..."

### Table 1: Upper bound for the required number of GC increments per block allocated as a function of the upper bound of the reachable memory.

"... In PAGE 4: ... This guarantees that a cycle finishes at the latest after the number of increments performed exceeds the amount of allocated memory, as is required by this approach (see [22] for the details). Table1 illustrates the worst- case number of increments that are required for the allocation of one block, as a function of the upper bound of reachable memory k = K / M. On the PowerPC processor, a marking or sweeping of one block takes about winc = 80 (compiled) machine instructions.... ..."

### Table 2: New bounds for fully dynamic reachability problems on directed acyclic graphs.

2005

Cited by 2