### Table 19. The architectural assumptions of the middle-tier, COM and their minimal common upper element (the resulting architecture).

"... In PAGE 13: ...Table19 . The architectural assumptions of the middle-tier, COM and their minimal common upper element (the resulting architecture).... In PAGE 126: ... COM. From Table19 - The architectural assumptions of the middle-tier, COM and their minimal common upper element (the resulting architecture).-, we can see that the baseline middle-tier architecture and the COM architecture differs in two aspects, i.... ..."

### Table 17. The architectural assumptions of the data-tier, the Clustra DBMS and their minimal common upper element (the resulting architecture).

"... In PAGE 12: ...able 16. Measured values for client-tier COTS products. ............................................................... 94 Table17 . The architectural assumptions of the data-tier, the Clustra DBMS and their minimal common upper element (the resulting architecture).... ..."

### Table 6: Threats in the Multiple Mix Protection Profile The list of related assumptions follows. Some of the assumptions are stated only to simplify the PPs, like the A.DedicatedHost, which excludes other processes on the same host of each mix, and are really not essential. However, there are assumptions, like the A.MinimalTrust, which are very important, because theystate explicitlywhen the entire mix network fails.

### Table 1: A single-fault focus provides a 40-fold speedup when using seven models per component. The Labels column gives the maximum and average number of environments in a label; Nogoods gives the count of the minimal nogoods, as well as the maximum and average number of assumptions in a nogood.

"... In PAGE 10: ... A reasonable diagnosis might be fzero(M1), left(M2)g or fleft(M2), right(A1)g. Table1 summarizes the results of using a single-fault focus. There is a dramatic 40-fold speedup in execution time when using the focus! There are several interre- lated reasons for this tremendous reduction.... ..."

### Table 4: Experimental results for crosstalk minimization.

"... In PAGE 13: ... Under this assumption, the optimization of cost function (3) is equal to the minimization of the maximum coupling capacitance in the channel, and maximizing the total noise slack is equal to minimizing the total coupling capacitance. The results under this noise slack assumption are shown in Table4 . In this table columns two and three include the values of the worst coupling capacitance and the total coupling capacitance in the channel, respectively, in the original layouts, while their corresponding values for the modi ed layouts are shown... In PAGE 14: ... As mentioned before we use the overlap length between adja- cent wires to represent the crosstalk between them. From Table4 we can see that an average of 16.4% reduction in maximum crosstalk can be achieved by our algorithm.... ..."

### Table 2: Efficiency-Assumptions in component-oriented diagnosis.

"... In PAGE 11: ... All these assumptions are introduced to reduce the computational effort required to solve the problem. Table2 provides an explanation of the assumptions along with the role they play (function), the domain they are about (case data, domain knowledge or task), and some references where they are discussed in more detail. Table 2: Efficiency-Assumptions in component-oriented diagnosis.... In PAGE 12: ... cd amp; dk It is used in computing probabilities for hypotheses. [de Kleer amp; Williams, 1989] Table2 : Efficiency-Assumptions in component-oriented diagnosis. name explanation is abouta function some references Assumptions for Efficiency minimality single or N-fault ingnorance of abnormal behaviour independence of hypothes search control knowledge existence of a hierarchical-layered device model existence of a hierarchical-layered behavioural model independence of fault probabilities Fig.... ..."

### lable. We are now in a position to state the solution to the modi ed H1 control problem formulated above, directly from Ba sar and Bernhard [3], Chap. 6. Proposition 1 Let gt; 0 be xed, and assumption A1 hold. Then, ( ) de ned above is nite, and for all gt; ( ): 1) There exists a minimal positive de nite solution, M , to

1995

Cited by 28

### Table 2: Minimal Starting Times of a Fine-Grained Communication Structure Algorithm 5.2 Compute the sequence of vertices v of a communication structures G in the traces of a schedule of G.

"... In PAGE 10: ... This is a contradiction to the assumption of this case. Table2 shows the minimal starting times for each vertex in the communication structure shown in gure 2. We assumed that for each vertex the computation time is C = 1, the latency L = 1, the overhead o = 1 and the gap g = 2.... ..."

### Table1: WiM results on minimal example sets

"... In PAGE 5: ... It results in japaneseCar(X):-producer(X,Y),place(Y,Z),isCountryJapan(Z). Again the translater is called to translate this Horn clause into F-logic and we receive japaneseCar:X lt;= car:X [producer- gt;F[place- gt;P[country- gt; apos;Japan apos;] 6 Discussion of results WiM ?D was examined on the following class and attribute de nitions: japaneseCar:X lt;= car:X [producer- gt;F[place- gt;P[country- gt; apos;Japan apos;] isMother:M lt;= person:M [son- gt;S] amp; not person:S[father- gt;X] factory:X lt;= carFactory:X factory:X lt;= aircraftFactory:X person:X lt;= child:X person:X lt;= adult:X eVehicle:X lt;= car:X[power- gt;electricity] eVehicle:X lt;= publicTransportVehicle:X[power- gt;electricity] family:F[husband- gt;H, wife- gt;W] lt;= person:H [spouse- gt;W] amp; person:W [spouse- gt;H] person:X[managed- gt;Y] lt;= person:X[boss- gt;Y] person:X[managed- gt;Y] lt;= person:X[boss- gt;Z] amp; person:Z[managed- gt;Y] person:X[mother- gt;M] lt;= not(person:X[father- gt;M]) amp; person:M[son- gt;X] person:X[mother- gt;M] lt;= not(person:X[father- gt;M]) amp; person:M[daughter- gt;X] In Table1 . there are numbers of both positive and negative examples needed for each new class/attribute described above.... In PAGE 6: ...No. of positive examples negative examples japaneseCar 1 2(cwa) isMother 3 9(cwa) factory 2 1(a) person 2 1(a) eVehicle 2 2(cwa) family 2 4(cwa) managed 4 6(cwa) mother 5 7(cwa) Table1 : Summary of WiM ?D results WiM ?D needed from 1 to 5 positive instances(objects) of classes. Negative examples were generated using close world assumption ( apos;cwa apos; in the 2nd column) or were found by as the assumption by the WiM?D itself ( apos;a apos;).... ..."