### Table 2: Comparative studies for ssVI and VI on the standard test-bed

"... In PAGE 19: ... In the literature, the eight problems are commonly referred to as 4x3CO, Cheese, 4x4, Part Painting, Tiger, Shuttle, Network, and Aircraft. Table2 presents detailed 6.... ..."

### Table 1: Testbed characteristics

2008

"... In PAGE 8: ... Moreover, we discarded some instances with numerical instability and which, after the variables were partitioned, were too easy to solve even by the classical Benders method 2. Table1 shows our nal testbed with the main characteristics of each instance. Standard variable partitioning has been applied|integer (and binary) variables are viewed as master variables x, and the continuous variables are viewed as slave variables y.... ..."

### TABLE testbed.

### Table 6 depicts the growth steps of the target dictionaries for the entire bootstrapping process (new entries in comparison to each previous step, , are in brackets). In the first run, for Spanish, 3,587 new subwords are added to the dictionary which leads to a size of 6,817, including those lexical entries already generated by the cognate identification routines (cf. Table 5). For Swedish, only 759 new subwords were generated in the first step. Remarkably, these entries lead to the acquisition of 1,361 new subwords in the next step. After 14 cycles, learning activities calm down with 7,154 subwords generated for Spanish, while after 9 runs 4,148 dictionary entries for Swedish are acquired.

"... In PAGE 4: ... Table6 : Dictionary Growth Steps 5. CLIR EXPERIMENTAL FRAMEWORK Our experiments were run on the OHSUMED corpus [8], which constitutes one of the standard IR testbeds for the medical domain.... ..."

### Table 4.1 summarizes some statistics for the testbed. Experiments were run on two sets of queries. The first set of queries came from the title fields of TREC topics 451-550 used for TREC-8 and TREC-9 Web Tracks. The standard TREC

2004

### Table 2: Average energy consumption (in Joule) during the setup operations in the multi-hop Mica2 testbed.

2005

"... In PAGE 11: ... 4.4 Energy Efficiency Table2 shows the itemized energy cost of the Z-MAC setup phase operations in the multi-hop benchmark. We run the setup phase for 30 times and report the average val- ues and standard deviations.... ..."

Cited by 50

### Table 1: World-Wide Grid testbed resources simulated using GridSim.

in A Deadline and Budget Constrained Cost-Time Optimisation Algorithm for Scheduling Task Farming . . .

2002

"... In PAGE 3: ... To enable the users to model and express their application processing requirements in terms of MI (million instructions) or MIPS (million instructions per second) on the standard machine, we assume the MIPS rating of PEs is same as the SPEC rating. Table1 shows the characteristics of resources ... ..."

Cited by 21

### Table 1: Comparison of speed of original deterministic algorithms and randomized versions on test-bed problems.

1998

"... In PAGE 2: ...s (N=2 (N 1))!, i.e., the search space size grows as the factorial of the square of N=2. Published algorithms for this problem all scale poorly, and the times for our deterministic solver (as shown in Table1 ) are among the best (see also Gomes et al. 1998b).... In PAGE 5: ... In addi- tion, a very low cutoff value can also be used to exploit the heavy-tails to the left of the median, and will allow us to solve previously unsolved problem instances after a suffi- cient number of restarts. In Table1 , the mean solution times in the Randomized column are based on empirically de- termined near-optimal cutoff values. For each randomized solution time the standard deviation is of the same order of magnitude as the mean.... In PAGE 6: ... Our fast restart strategy exploits this. See Table1 for other improvements due to randomiza- tion. Until now, the 3bit-adder problems had not been solved by any backtrack-style procedure.... ..."

Cited by 230

### Table 1: Comparison of speed of original deterministic algorithms and randomized versions on test-bed problems.

1998

"... In PAGE 2: ... i.e., the search space size grows as the factorial of the square of a6a12a11a14a13 . Published algorithms for this problem all scale poorly, and the times for our deterministic solver (as shown in Table1 ) are among the best (see also Gomes et al. 1998b).... In PAGE 5: ... In addi- tion, a very low cutoff value can also be used to exploit the heavy-tails to the left of the median, and will allow us to solve previously unsolved problem instances after a suffi- cient number of restarts. In Table1 , the mean solution times in the Randomized column are based on empirically de- termined near-optimal cutoff values. For each randomized solution time the standard deviation is of the same order of magnitude as the mean.... In PAGE 6: ... Our fast restart strategy exploits this. See Table1 for other improvements due to randomiza- tion. Until now, the 3bit-adder problems had not been solved by any backtrack-style procedure.... ..."

Cited by 230

### Table 1: World-Wide Grid testbed resources simulated using GridSim.

2002

"... In PAGE 3: ... To enable the users to model and express their application processing requirements in terms of MI (million instructions) or MIPS (million instructions per second) on the standard machine, we assume the MIPS rating of PEs is same as the SPEC rating. Table1 shows the characteristics of resources simulated and their PE cost per time unit in G$ (Grid dollar). The simulated resources resemble the WWG ... ..."

Cited by 21