Results 1 - 10 of 9,820

Table 4. Algorithm schedule.

in Automated Placement and Routing of Cell Matrix Circuits
by Donald H. Cooley, Thomas Kent
"... In PAGE 37: ...Table4 consistently yielded fast placement and routing of netlists with complexity levels varying from a few gates to thousands of gates. Table 4.... In PAGE 46: ... Through experimentation, I found several schedules resulting in fast and efficient placement and routing of a wide variety of circuit sizes (up to 2000 gates). Figure 7 illustrates the routing iterations of a 100-gate synthetic circuit with 230 wires on a 30x30 Cell Matrix with a 3 percent bad cell rate using the schedule in Table4 . plots the percentage of routed wires as a function of iteration number averaged over ten runs, with netlists and bad cell maps of similar complexities generated randomly for each run.... ..."

Table 4: Costs associated with each possibility. Positive and negative values express failure and success respectively. Observing time is lost when an observation requiring good seeing is scheduled during a bad or moderate actual seeing night. Conversely, observing time is ine ciently used when an observation with no speci c seeing requirements is scheduled during a good seeing night. The results obtained for both prediction sets are cross-tabulated in Tables 6 to 14. Various measures of interest can be derived from these contingency tables. They are described below in a convenient way if one wishes to diminish the number of \false alarms quot;, i.e. bad seeing when good seeing is expected, or to minimize a given cost function, etc.

in Dynamical Recurrent Neural Networks and Pattern Recognition Methods for Time Series Prediction: Application to Seeing and Temperature Forecasting in the Context of ESO's VLT Astronomical Weather Station
by Alex Aussem, Fionn Murtagh, Alex Aussem Yz, Alex Aussem Yz, Fionn Murtagh \upsilon, Fionn Murtagh \upsilon, Marc Sarazin, Marc Sarazin Y

Table 19: Effect of Scheduling Schemes - CPU Times

in Parameterized Heuristics for Project Scheduling - Biased Random Sampling Methods
by Andreas Schirmer, Sven Riesenberg
"... In PAGE 5: ...able 18: Effect of Scheduling Schemes and Iterations - Deviations...........................................................28 Table19 : Effect of Scheduling Schemes - CPU Times.... In PAGE 35: ... Indeed, for the bad rules the PSS re- mains dominant even after 100 iterations, regardless of the RSS used. The bearing of the SS on efficiency is shown in Table19 where the CPU times for the SS-RSS combinations are averaged over all priority rules. In order to control for the influence of differ-... ..."

Table 2: Average deviations from best solution | n = 60 only project scheduling knowledge is contained in the SGS (due to this knowledge, the results are not as bad as one may expect). These random procedures serve as benchmarks as they allow us to evaluate how much the results can be improved by incorporating more project scheduling knowledge. Employing random sampling methods as benchmark solutions is common for the evaluation of scheduling heuristics (cf., e.g., Conway et al. [8]). Generally, any procedure should perform considerably better than a pure random procedure, especially if the same SGS is used. The results for the j30 set show, however, that two of the genetic algorithms yield a higher average deviation from the optimal makespan than the respective random sampling procedure with the same SGS. For the two other instance sets the two random sampling procedures perform worst among all heuristics if 5000 iterations are considered.

in Experimental Evaluation of State-of-the-Art Heuristics for the Resourc-Constrained Project Scheduling Problem
by Sönke Hartmann, Rainer Kolisch 2000
Cited by 28

Table 12: Effect of Scheduling Schemes - CPU Times

in Advanced Biased Random Sampling in Serial and Parallel Scheduling
by Andreas Schirmer 1997
"... In PAGE 23: ... Indeed, for the bad rules the PSS remains dominant even after 100 iterations, regardless of the RSS used. The bearing of the SS on efficiency is shown in Table12 where the CPU times for the SS-RSS combinations are averaged over all priority rules. In order to control for the influence of differ-... ..."
Cited by 5

Table 12: Effect of Scheduling Schemes - CPU Times

in Advanced Biased Random Sampling in Serial and Parallel Scheduling
by Andreas Schirmer
"... In PAGE 24: ... Indeed, for the bad rules the PSS remains dominant even after 100 iterations, regardless of the RSS used. The bearing of the SS on efficiency is shown in Table12 where the CPU times for the SS-RSS combinations are averaged over all priority rules. In order to control for the influence of differ-... ..."

Table 3. Distribution of 20 overcover links in 160 minimum shift schedules

in Driver Scheduling using Genetic Algorithms with Embedded Combinatorial Traits
by Ann S.K. Kwan, Raymond S.K. Kwan, Anthony Wren 1999
"... In PAGE 17: ... In fact often only one of the pair of shifts is at fault. This explains why some of the overcover links might not be bad as indicated in Table3 . Nevertheless, Table 3 shows that 9 out of 20 overcover links are most likely to be bad, and the banning of another 8 of them probably would not be detrimental.... ..."
Cited by 8

TABLES/bad_nets :\

in 1. Introduction Exim configuration at the University of Cambridge
by Tony Finch

Table 6: Approaches for tackling Coverage Constraints

in The State of the Art of Nurse Rostering
by Edmund Burke, Patrick De Causmaecker, Greet Vanden Berghe, Hendrik Van Landeghem 2004
"... In PAGE 35: ... They either allow all the possible variations of schedules and penalise the bad ones, or they select schedules from a limited set of desired shift sequences. Many researchers are aware of the necessity of small violations of the coverage constraints when it comes to scheduling in practice (see Table6 ), and penalise them in a cost function. The approaches that are grouped in the first column (where understaffing is allowed) and in the third one (overstaffing is allowed), take coverage as a goal to be reached.... ..."
Cited by 24

Table 2. Number of cycles to compute expression trees using: Right-Left, Left-Right and OptSchedule

in Code Generation for Fixed-Point DSPs
by Guido Araujo, Sharad Malik 1998
"... In PAGE 21: ... The metric used to compare the code was the number of cycles that takes to compute the expression tree. Observe from Table2 that algorithm OptSchedule produces the best code when compared with two SC schedules, what is expected since we have proved its opti- mality. Notice that although SC schedules can sometimes produce optimal code, it can also generate bad quality code, as it is the case for expression tree 6.... ..."
Cited by 5
Results 1 - 10 of 9,820