"... In PAGE 13: ... Consequently, they will be scheduled using XY routing and will be in level 2. The fourth column of Table5 shows the communication schedule obtained from using F CF S with re-routing.... In PAGE 14: ... This was done by starting with the messages in the highest level and then trying to re-route them to see if any improvement can be obtained. In the example, and also shown in column 3 of Table5 , message 4 is in level 3 and is the rst candidate for rescheduling. As previ- ously described, a message must satisfy deadlock avoidance and other criteria before it can be rescheduled.... In PAGE 14: ...ove up to level 1 in the schedule. Figure 7 shows message 1 being re-routed. At this point, this technique stops since the messages in level one obviously cannot be re-routed to move to a lower level. Column 5 of Table5 shows the communication schedule obtained from using MISCOM with re-routing. The level sum has improved to 8 from the previous value of 11.... In PAGE 15: ... This new algorithm takes the output of the rescheduled F CF S and uses it as input to the rescheduled MISCOM technique. Thus, a CG is built from the message list corresponding to column 4 of Table5 . The MISCOM technique is applied which yields level 1 nodes of (5, 6, 7, 3 apos;, 4 apos;).... In PAGE 15: ... Only message 1 can be re-routed to achieve a better performance. The resultant schedule is shown in the last column of Table5 . It is slightly di erent from the rescheduled MISCOM output, but has the same sum or overall processing time value.... ..."

"... In PAGE 9: ...Table 3. Generation rescheduling results of the 39-bus system Note that, in Table3 , actual P margin refers to the power margin acquired using transient stability program (time domain simulation), and estimated P margin is the power margin obtained directly from the optimization. The results show that all methods can improve the stability margin of the base case, and the better ones result from the proposed optimization approach.... In PAGE 9: ... The results show that all methods can improve the stability margin of the base case, and the better ones result from the proposed optimization approach. From the results shown in Table3 , although there is a mismatch between actual margin, and estimated margin but the generation power output according to estimated margin does give a direction how to redispatch the generation units in order to improve the stability limit. ... ..."

"... In PAGE 7: ... Figure 4(b) shows the messages after re-routing has been applied with dashed lines denoting that messages 3 and 4 have been re-routed. The fourth column of Table4 shows the resulting communication schedule with re-routed messages denoted with a apos; mark.... In PAGE 7: ... This is done by starting with the messages in the highest level and trying to re-route them to see if any improvement can be obtained. In the example, and also shown in column 3 of Table4 , message 4 is in level 3 and is the first candidate for rescheduling. This message is eligible for re-routing and doing so moves it up to level... In PAGE 8: ...Table 4: Level assignments with rescheduling one cannot be moved to a lower level. Column 5 of Table4 shows the communication schedule obtained and the level sum has improved to 8 from the previous value of 11. Thus, a 27.... ..."

"... In PAGE 18: ...actors weighted at 35%, 15%, and 6%, respectively (http://store.eiu.com). Table20 examines the 27 studies concerned with debt rescheduling, in which three types of variables were used, namely economic, financial and political. The economic and financial variables were used in each of the 27 studies, whereas political variables were used in only 9 studies.... ..."

"... In PAGE 5: ... We have used the same eight solvents as probes to map the entire surface of HEWL. Table1 shows the clusters with the lowest Boltzmann average free energies, including size, probability, average free en- ergy, average electrostatic energy, van der Waals and desolvation free energy contributions, the number of the subclusters in each cluster, the distance between the hub of the cluster and any residue in site C (N59, W62, W63 and A107), and the residue closest to the cluster (not necessarily in site C). The lowest free energy clusters were found to be in site C of HEWL for all solvents, although DMSO and methyl- ene chloride are not as deeply inside the pocket as the other six ligands.... ..."

"... In PAGE 8: ... Fixed strategies such as always rejecting or always bumping fail to be effec- tive. Table1 shows a comparison of the average performance of the two fixed strategies. (The exact experimental set-up is described in more detail in Sec- tion 5.... In PAGE 13: ... This graph shows that the use of our scheduling difficulty model is able to reduce scheduling failures. Also, the high failure rate caused by uncontrolled cascad- ing of bumps, as we saw in Table1 for the Always-bump strategy, is avoided. Next, we evaluate the effect of varying our scheduling difficulty model in the busy/non-busy hierarchy.... ..."

"... In PAGE 8: ... Fixed strategies such as always rejecting or always bumping fail to be effec- tive. Table1 shows a comparison of the average performance of the two fixed strategies. (The exact experimental set-up is described in more detail in Sec- tion 5.... In PAGE 13: ... This graph shows that the use of our scheduling difficulty model is able to reduce scheduling failures. Also, the high failure rate caused by uncontrolled cascad- ing of bumps, as we saw in Table1 for the Always-bump strategy, is avoided. Next, we evaluate the effect of varying our scheduling difficulty model in the busy/non-busy hierarchy.... ..."

"... In PAGE 18: ... The economic and financial variables were used in each of the 27 studies, whereas political variables were used in only 9 studies. Table21 presents the number of variables used in debt rescheduling, as well as their frequency. All three variables have been used in 9 studies, two of the three variables were used in the remaining 18 studies, and no study used only one of the three variables.... ..."

"... In PAGE 16: ... Thus, the idle time of each resource is minimized. Table13 shows that convergence to the optimum solution with respect to concurrent rescheduling0task completion is reached after four iterations. That is, the scheduler should reschedule 60 tasks, estimated to take approximately 20 s ~i.... ..."

"... In PAGE 16: ...http://store.eiu.com). Table20 examines the 27 studies concerned with debt rescheduling, in which three types of variables were used, namely economic, financial and political. The economic and financial variables were used in each of the 27 studies, whereas political variables were used in only 9 studies.... ..."